Eric Goldman’s Techdirt Profile


About Eric Goldman

I am a Professor at Santa Clara University School of Law.

Posted on Techdirt - 29 September 2021 @ 6:23am

The SHOP SAFE Act Is A Terrible Bill That Will Eliminate Online Marketplaces

from the don't-hold-back-now dept

We've already posted Mike's post about the problems with the SHOP SAFE Act that is getting marked up today, as well as Cathy's lamenting the lack of Congressional concern for what they're damaging, but Prof. Eric Goldman wrote such a thorough and complete breakdown of the problems with the bill that we decided that was worth posting too.

[Note: this blog post covers Rep. Nadler’s manager’s amendment for the SHOP SAFE Act, which I think will be the basis of a committee markup hearing today. If Congress were well-functioning, draft bills going into markup would be circulated a reasonable time before the hearing, so that we can properly analyze them on a non-rush basis, and clearly marked as the discussion version so that we’re not confused by which version is actually the current text.]

The SHOP SAFE Act seeks to curb harmful counterfeit items sold through online marketplaces. That’s a laudable goal that I expect everyone supports. However, this bill is itself a giant counterfeit. It claims to focus on “counterfeits” that could harm consumer “health and safety,” but those are both lies designed to make the bill seem narrower and more balanced than it actually is.

Instead of protecting consumers, this bill gives trademark owners absolute control over online marketplaces by overturning Tiffany v. eBay. It creates a new statutory species of contributory trademark liability that applies to online marketplaces (defined more broadly than you think) selling third-party items that bear counterfeit marks and implicate “health and safety” (defined more broadly than you think), unless the online marketplace operator does the impossible and successfully navigates over a dozen onerous and expensive compliance obligations.

Because the bill makes it impossible for online marketplaces to avoid contributory trademark liability, this bill will drive most or all online marketplaces out of the industry. (Another possibility is that Amazon will be the only player able to comply with the law, in which case the law entrenches an insurmountable competitive moat around Amazon’s marketplace). If you want online marketplaces gone, you might view this as a good outcome. For the rest of us, the SHOP SAFE Act will reduce our marketplace choices, and increase our costs, during a pandemic shutdown when online commerce has become even more crucial. In other words, the law will produce outcomes that are the direct opposite of what we want from Congress.

In addition to destroying online marketplaces, this bill provides the template for how rightsowners want to reform the DMCA online safe harbor to make it functionally impossible to qualify for as well. In this respect, the SHOP SAFE Act portends how Congress will accelerate the end of the Web 2.0 era of user-generated content.

[The rest of this post is 4k+ words explaining what the bill does and why it sucks. You might stop reading here if you don’t want the gory/nerdy details.]

Who’s Covered by the Bill

The bill defines an “electronic commerce platform” as “any electronically accessed platform that includes publicly interactive features that allow for arranging the sale or purchase of goods, or that enables a person other than an operator of the platform to sell or offer to sell physical goods to consumers located in the United States.”

Clearly, the second part of that definition targets Amazon and other major marketplaces, such as eBay, Walmart Marketplace, and Etsy. I presume it also includes print-on-demand vendors that enable users to upload images, such as CafePress, Zazzle, and Redbubble (unless those vendors are considered to be retailers, not online marketplaces).

The first part of the definition includes services with “publicly interactive features that allow for arranging the sale or purchase of goods.” This is a bizarre way to describe any online marketplace, and it covers something other than enabling third-party sellers (that’s the second part of the definition), so what services does this describe? Read literally, all advertising “allow[s] for arranging the sale or purchase of goods,” so this law potentially obligates every ad-supported publisher to undertake the content moderation obligations the bill imposes on online marketplaces. That doesn’t make sense, because the bill uses the undefined term “listing” 11 times, and display advertising isn’t normally considered to be a listing. Still, this wording is unusual and broad — and you better believe trademark owners like its breadth. If the bill wasn’t meant to regulate all ads, the bill drafters should make that clear.

Like most Internet regulations nowadays, the bill distinguishes entities based on size. See my article with Jess Miers on how legislatures should do that properly. The bill applies to services that have “sales on the platform in the previous calendar year of not less than $500,000.” Some problems with this distinction:

  • The bill doesn’t define “platform,” so it’s unclear what revenues count. In Amazon’s case, is it only revenues from the marketplace or does it also include the revenues from Amazon’s retailing function? If the latter, then the definition will pick up smallish online retailers that have small marketplace components.
  • The bill also doesn’t distinguish between gross and net revenue. So, for example, assume a site takes a 10% commission on sales. If a service has $500k in merchandise sales (gross revenue), but only keeps $50k in commissions (net revenue), is it covered by the law or not? I think the bill covers gross revenue, which means the bill reaches companies with small net revenues.
  • As usual, the bill doesn’t provide a phase-in period. A service may not know its revenues until some time after the calendar year closed, but it would be obligated to comply with the law from the beginning of the calendar year. As usual, then, this forces services below the revenue threshold to comply anticipatorily in case they clear the threshold. How hard is it for bills to include a phase-in period?

I’d fret more about the $500k threshold, but it’s likely to be irrelevant anyways. The bill also applies to smaller services once they receive 10 NOCI notices over their lifetimes from all sources. (Unlike the other services, these services get a six-month phase-in period).

To qualify as a NOCI, the notice must (1) refer to the SHOP SAFE Act, (2) “include an explicit notification of the 10-notice limit and the requirement of the platform to publish” the NOCI disclosures below (I have no idea what this element means), and (3) “identify a listing on the platform that reasonably could be determined to have used a counterfeit mark in connection with the sale, offering for sale, distribution, or advertising of goods that implicate health and safety.” (So, a NOCI counts against the 10-notice threshold if it “reasonably could be determined” that the listing was counterfeit, even if the NOCI is actually wrong.)

A month after getting its first NOCI, the service must publicly post an attestation that it has less than $500k in revenue along with a running tally of the number of NOCIs received… I guess for shits and giggles so that trademark owners can compete to be the one to put the service over the 10 NOCI threshold? I mean, even tiny services will quickly accrue 10 NOCIs. Indeed, I imagine rightsowners will coordinate their NOCIs to ensure that small services clear this threshold and are obligated to comply with the law. Thus, the 10 lifetime NOCIs threshold is a ruse to mislead people that smaller services aren’t governed by the law, when of course they will be.

What’s Regulated?

The law applies to counterfeit “goods that implicate health and safety,” defined as “goods the use of which can lead to illness, disease, injury, serious adverse event, allergic reaction, or death if produced without compliance with all applicable Federal, State, and local health and safety regulations and industry-designated testing, safety, quality, certification, manufacturing, packaging, and labeling standards.” I mean, pretty much every physical product meets this definition, right? Virtually any poorly-designed or nonconforming physical item has the capacity to cause personal injury. For example, electronic items that don’t comply with industry standards can cause physical harm from electrical charges, which means every electronic item is categorically within the bill’s scope even if the allegedly counterfeited item actually complies with industry standards. Now, replicate that analysis for other goods and tell me which categories of goods lack the capacity to cause harm. Once again, the “health and safety” framing is another deceptive ruse because the bill functionally applies to all goods, not just especially risky goods.

Overturning Tiffany v. eBay

In 2010, the Second Circuit issued a watershed decision about secondary trademark infringement. Essentially, the court held that eBay wasn’t liable for counterfeit sales of Tiffany items because eBay honored takedown notices and Tiffany’s claims sought to hold eBay accountable for generalized knowledge. That ruling has produced a kind of détente in the online secondary trademark infringement field, where we just don’t see broad counterfeiting lawsuits against online marketplaces any more.

The SHOP SAFE Act ends that détente. First, it creates a new statutory contributory trademark infringement claim for selling the regulated items. Second, the bill says that the new contributory claim doesn’t preempt other plaintiff claims, so trademark owners will still bring the standard statutory direct trademark infringement claim and common law contributory trademark claims (and dilution, false designation of origin, etc.). Third, online marketplaces nominally can try to “earn” a safe harbor from the new statutory contributory liability claim (but not from the other legal claims) by jumping through an onerous gauntlet of responsibilities. Those requirements will impose huge compliance costs, but those investments won’t prevent online marketplaces from being dragged into extraordinarily expensive and high-stakes litigation over eligibility for this defense. Fourth, the law imposes a proactive screening obligation, something that Tiffany v. eBay rejected. Fifth, unlike Tiffany v. eBay, generalized knowledge can create liability, and takedown notices aren’t required as a prerequisite to liability. Sixth, in litigation over direct trademark infringement and common law contributory trademark infringement claims, trademark owners can cite compliance/non-compliance with the defense factors against the online marketplace, putting the online marketplace in a worse legal position than they currently are in.

All told, the SHOP SAFE Act will functionally repeal the Tiffany v. eBay standard that has fostered the growth of online marketplaces for the last decade-plus, and usher in a new era of online shopping that will likely exclude online marketplaces entirely.

The “Safe Harbor” Preconditions

To earn protection from the newly created contributory trademark infringement doctrine, online marketplaces must perfectly implement all of the following 13 requirements:

1. Determine, and periodically confirm, that third-party sellers have a registered US agent for service or a designated “verified” US address for service. (Just wait until other countries require the equivalent from US-based online sellers on foreign marketplaces. A new frontier for a trade war.)

2. Verify the third-party seller’s identity, principal place of business, and contact information through “reliable documentation, including to the extent possible some form of government-issued identification.” (What is “reliable” documentation, and how much risk will online marketplaces be willing to take?)

3. Require the third-party seller to take reasonable steps to verify the authenticity of its goods and attest to those steps. This requirement doesn’t apply to sellers who sell less than $5k/yr and lists no more than five of the same items per year. (Is the online marketplace liable if the seller doesn’t actually take reasonable steps? How can the online marketplace “require” independent sellers to do this?)

4. Impose TOS terms that the third-party seller (1) won’t use counterfeit marks, (2) consents to US jurisdiction, and (3) designates a US agent for service or has a verified US address for service. (Can trademark owners take advantage of the US jurisdiction consent between the online marketplace and its third-party sellers? Normally trademark owners aren’t third-party beneficiaries of that contract. Also, that consent isn’t limited to jurisdiction over counterfeit claims — it’s over everything the TOS might govern.)

5. Conspicuously display on the platform:

  • the third-party seller’s verified principal place of business,
  • contact information,
  • identity of the third-party seller, and
  • the country from which the goods were originally shipped from the third-party seller

But the online marketplace isn’t required to display “the personal identity of an individual, a residential street address, or personal contact information of an individual, and in such cases shall instead provide alternative, verified means of contacting the third-party seller.”

6. Conspicuously display “in each listing the country of origin and manufacture of the goods as identified by the third-party seller, unless such information was not reasonably available to the third-party seller and the third-party seller has identified to the platform the steps it undertook to identify the country of origin and manufacture of the goods and the reasons it was unable to identify the same.” This requirement doesn’t apply to sellers who sell less than $5k/yr and lists no more than five of the same items per year.

7. Require third-party sellers to “use images that accurately depict the goods sold, offered for sale, distributed, or advertised on the platform.” (Does this create an affirmative obligation to include images? While rare, I believe that some marketplace sellers sometimes currently sell their items without including any photo. Also, product shots have been a constant source of copyright litigation. The manufacturer can sue the seller for copying its shots; the manufacturer can sue for false advertising if non-official shots aren’t “accurate,” and freelancers love to sue over product shots they took and ones they think are too similar to the ones they took.)

8. Undertake “reasonable proactive measures for screening goods before displaying the goods to the public to prevent the use by any third-party seller of a counterfeit mark in connection with the sale, offering for sale, distribution, or advertising of goods on the platform. The determination of whether proactive measures are reasonable shall consider the size and resources of a platform, the available technological and non-technological solutions at the time of screening, the information provided by the registrant to the platform, and any other factor considered relevant by a court.” (This is the most coveted payload for trademark owners. Every rightsowner wants UGC services to engage in proactive screening. The screening won’t be limited to harmful counterfeit goods, and consider how courts will punish online marketplaces for undertaking this proactive screening in their analysis of direct and contributory trademark infringement.)

9. Provide “reasonably accessible electronic means by which a registrant and consumer can notify the platform of suspected use of a counterfeit mark.” (What are the odds that the consumer notifications will be made in good faith? Consider, in particular, how a dissatisfied buyer could weaponize this provision for reasons having nothing to do with counterfeiting. Note also how buyer complaints of counterfeiting, when not accurate — and buyers won’t necessarily know — could create scienter on the online marketplace’s part, and the countermoves by the marketplace could work to the detriment of the marketplace, the seller, AND the manufacturer by reducing their online marketplace sales.)

10. Implement “a program to expeditiously disable or remove from the platform any listing for which a platform has reasonable awareness of use of a counterfeit mark in connection with the sale, offering for sale, distribution, or advertising of goods.” The online marketplace’s scienter may be inferred from:

  • information regarding the use of a counterfeit mark on the platform generally,
  • general information about the third-party seller,
  • identifying characteristics of a particular listing, or
  • other circumstances as appropriate.

(This differs from the DMCA online safe harbor in many ways. The most obvious is that online marketplaces can be liable for the new statutory contributory trademark claim even if trademark owners never send them takedown notices. Among other things, this factor also emboldens trademark owners to send notices like “there are counterfeits on your site — find and remove them” without identifying any specific infringing listing. It seems those generalized notices would confer scienter sufficient to impose contributory trademark infringement. This, of course, directly rejects the Tiffany v. eBay precedent, which said such generalized knowledge wasn’t enough.)

An online marketplace can restore a listing “if, after an investigation, the platform reasonably determines that a counterfeit mark was not used in the listing.” (How many services will want to do the investigation, and how confident will the service be that the trademark owner will agree that they “reasonably” determined the listing wasn’t counterfeit? In practice, once a listing is down, it ain’t going back up.)

11. Implement “a publicly available, written policy that requires termination of a third-party seller that reasonably has been determined to have engaged in repeated use of a counterfeit mark.” (Note how this combines several parts of the DMCA online safe harbor, including the obligation to adopt a repeat infringer policy, to publish the repeat infringer policy, and to reasonably implement the repeat infringer policy.)

Apparently online marketplaces are free to create their own repeat termination policy, but the bill says “Use of a counterfeit mark by a third-party seller in 3 separate listings within 1 year typically shall be considered repeated use.” (This sidesteps the obvious question of how services “know” that a seller used the counterfeit mark. Remember, in obligation #10, online marketplaces must terminate listings when the service has a “reasonable awareness,” which isn’t conclusive proof that counterfeiting actually took place. So does each removal based on that lowered scienter count as one of the three strikes?)

Online marketplaces can reinstate terminated sellers in some circumstances, none of which have any realistic chance of happening.

12. Take reasonable measures to ensure terminated sellers don’t reregister on the service. (Another coveted item by rightsowner: a permanent staydown.)

13. Provide “a verified basis to contact a third-party seller upon request by a registrant that has a bona fide belief that the seller has used a counterfeit mark.” (I didn’t understand this provision because the trademark owners should already have all of the information they need to blast counterfeiters from obligation #5).

Whew! Could trademark owners ask for anything more? These obligations are pretty much their dream wishlist.

Liability for Bogus NOCIs

The bill creates a new cause of action for bogus takedown notices sent to online marketplaces. I’m going to dig into this cause of action, but no need to master the details: Congress has learned absolutely nothing from the failure of 17 USC 512(f), so there’s no possible way for any plaintiff to benefit from this provision.

The cause of action: “Any person who knowingly makes any material misrepresentation in a notice to an electronic commerce platform that a counterfeit mark was used in a listing by a third party seller for goods that implicate health and safety shall be liable in a civil action for damages by the third-party seller that is injured by such misrepresentation, as the result of the electronic commerce platform relying upon such misrepresentation to remove or disable access to the listing, including temporary removal or disablement.” If the third-party seller declines to sue the trademark owner, the online marketplace can sue (with the third-party seller’s consent) if the trademark owner sent 10+ bogus notices. The bill provides statutory damages that range between $2,500-$75,000 per notice.

That sounds swell, but it’s useless for two reasons.

First, the bill doesn’t require trademark owners to send takedown notices in the first place. Trademark owners can sue online marketplaces for contributory trademark infringement without ever sending a takedown notice. So if trademark owners face potential liability for sending bogus takedown notices, why send them at all? Or trademark owners will send very generalized notices that don’t trigger liability for them but will trigger liability for the online marketplace.

Second, and more importantly, the cause of action requires a scienter that plaintiffs can’t prove. How can a third-party seller or online marketplace show the trademark owner knowingly made a material misrepresentation in their takedown notices? They can’t — unless they find smoking-gun evidence in discovery, but their complaints won’t survive a motion to dismiss sufficient to get to discovery. So there’s no way to win.

The “knowingly makes any material misrepresentation” standard is virtually identical to the 512(f) standard (“knowingly materially misrepresents”), so I expect courts will interpret the scienter standards the same. The Ninth Circuit killed 512(f) claims when it concluded in the Rossi case that the copyright owner’s subjective belief of infringement was good enough to defeat liability. As a result, over the past 20+ years, there has been only a small handful of 512(f) cases that have led to damages, and those few mostly involve default judgments. If trademark owners similarly can defend against this claim based on their subjective belief that counterfeiting is taking place, plaintiffs cannot win.

This provision is yet another ruse. It’s designed to make people think there’s a disincentive against trademark owner overclaims; but anyone who knows the 512(f) caselaw knows that this cause of action is completely worthless and a waste of everyone’s time.

Selected Problems with the Bill

What is “Counterfeiting”? The bill defines “counterfeit mark” as “a counterfeit of a mark” (I can’t make this up). But there’s actually a lot of confusion about what constitutes counterfeiting. See, e.g., my post about the trademark enforcements involving the “EMOJI” word mark, where they take the position that a marketplace item using the term “emoji” in the product name or description “counterfeits” their mark (seriously, look at the example from their exhibit and tell them that’s not bogus). A similar issue arises with print-on-demand services, where trademark owners take the position that any variation of their mark being manufactured onto a good constitutes counterfeiting, even if it’s parodic or an obvious joke. Thus, the bill’s grammar restricting the “use of counterfeit marks” potentially covers a much wider range of activity than classic piratical counterfeiting. Trademark owners will weaponize that ambiguity.

Lack of State Preemption. The Lanham Act doesn’t preempt state trademark laws, so this law isn’t likely to preempt any state law equivalents. It also would leave in place laws like the Arkansas Online Marketplace Consumer Inform Act, which has overlapping but different requirements than the SHOP SAFE Act. That overlap jacks up compliance costs and risks even more. While the SHOP SAFE Act is terrible and should never pass, it is even more terrible without a preemption provision.

Country of Origin Problems. The mandatory reporting of products’ country of origin is a liability trap. The bill excludes the smallest sellers from making this disclosure, but plenty of small-scale sellers will be obligated nonetheless, and they (and even bigger players) are sure to botch this because the law is confusing and the information won’t always be available to resellers. Any error on country-of-origin disclosures sets up the third-party sellers for false advertising claims. (Per Malwarebytes, the online marketplace should qualify for Section 230 protection for the Lanham Act false advertising claims). This gives trademark owners a second way of targeting third-party sellers: even if those sellers aren’t engaging in counterfeiting or any trademark infringement at all, country-of-origin false advertising claims can still be weaponized to drive them out of the marketplace.

Repudiation of the 512 Deal. The DMCA online safe harbor struck a grand bargain: online copyright enforcement would be a shared responsibility. Copyright owners would identify infringing items, and service providers would then remove those items. There has never been a trademark equivalent of the DMCA, but the Tiffany v. eBay case has de facto created a similar balance. Unsurprisingly, copyright owners hate the DMCA's shared responsibility, and they have tried to undermine that deal through lawfare in courts. Trademark owners similarly want a different deal.

This bill, as Congress’ first trademark complement to the DMCA, emphatically repudiates the DMCA deal. It gives trademark owners everything they could possibly want: turning online marketplaces into their trademark enforcement deputies, getting them to proactively screen for infringing items, making them wipe out listings without having to send listing-by-listing notices, upfront disclosure of the information needed to sue the sellers (rather than going through the 512(h) subpoena process), and permanent staydown of allegedly recidivist sellers.

Not only does this represent terrible trademark policy, but it’s a preview of how copyright owners will force DMCA safe harbor reform. They will want all of the same things: proactive monitoring of infringement, no need to send item-specific notices, authentication of users before they can upload, and staydown requirements. The SHOP SAFE Act isn’t just about counterfeits; it’s a proxy war for the next round of online copyright reform, and the open Internet doesn’t have a chance of surviving either reform.

“Reasonableness” Isn’t Reasonable to Online Marketplaces. I’ve blogged many times about how a “reasonableness” standard of liability in the online context is a fast-track to the end of UGC. As a legal standard, “reasonableness” often can’t be resolved on motions to dismiss because it’s fact-intensive and defendants can’t tell their side of the story at that procedural stage. As a result, “reasonableness” standards substantially increase the odds that lawsuits survive the motion to dismiss and get into discovery, which raises the defense costs by a factor of 10 or more.

The bill contains 21 instances of the term “reasonable” or variations. Each and every one of those is a fight the defendants can’t cost-justify. That means defendants will give up at the earliest opportunity or, more likely, self “censor” to avoid any potential courtroom battle over their “reasonableness.”

Too Many Defense Factors Makes the Defenses Unwinnable. More generally, to avoid the new cause of action, online marketplaces must win each and every one of the 13 preconditions (many of which have subparts). In other words, they must do everything perfectly AND prove all 13 elements to the court’s satisfaction. Safe harbors with that many prerequisites are extraordinarily costly because the plaintiffs can contest each element and engage in expensive discovery related to them. The DMCA online safe harbor has functionally failed for this reason: it’s too expensive for startups to prove they qualify, and copyright owners can weaponize those costs intentionally to drive entities out of the industry. This has turned the DMCA online safe harbor into a sport of kings, so only larger companies can afford it, which has exacerbated the concerns about “Big Tech” market consolidation. The SHOP SAFE Act replicates the structure that failed in the DMCA online safe harbor, so it’s predictable that the SHOP SAFE defenses also will fail to help out online marketplaces, leaving them highly vulnerable to the new cause of action.

Goodbye, Scalability. The Internet enables scalable operations in new and important ways. That scalability has created new functionality that never existed in the offline world — like online marketplaces. The SHOP SAFE Act blows scaling apart. Not only do the “reasonableness” requirements require careful attention to the facts, the bill makes it impossible to have true self-service signups of third-party sellers. Instead, there will need to be several levels of human review of new signups to satisfy the various authentication requirements. Furthermore, the proactive screening requirement will also require substantial human monitoring because determining “counterfeits” cannot be delegated solely to the machines. The absence of scalability and the need for substantial human labor will reward services that are really small, like a one-person operation, or really large, like a market-dominant player. Thus, SHOP SAFE’s elimination of scalability will exacerbate competition problems in the online retailing world.

Who Cares About Privacy? Trademark owners demanded the WHOIS system to make it easier for them to sue domain name registrants. The WHOIS system has collapsed due to the GDPR, which exposed how the WHOIS system was highly privacy-invasive. The SHOP SAFE Act doubles down on privacy invasions in two ways.

First, it requires online marketplaces to collect lots of sensitive information they don’t want, such as government-issued IDs. Those databases are honeypots for law enforcement and hackers.

Second, it requires publication of some information that sellers might consider private, especially if they are small operations with close identity between their professional and personal lives. (The bill’s exclusion of some private information incompletely addresses this concern.) For example, that information can be highly sensitive for sellers of controversial items who can be targeted by trolls and haters for local ostracism or physical attacks like swatting, and competitors can use this information too to engage in anti-competitive harassment.

Just like WHOIS struck a lopsided balance between trademark owners’ interests and registrant privacy, the SHOP SAFE Act similarly tosses privacy concerns under the trademark owners’ bus.

Why Would Anyone Support This Bill? This bill will kill online marketplaces and make markets less efficient. Where the online marketplace owner has a retailing function, like Amazon and Walmart, they can shut down the marketplace and subsume some items into their standard retailing function. That transition cuts off the long tail of items consumers expect to find online, and it burns hundreds of thousands of independent businesses that currently thrive in the marketplace system but become irrelevant in a retailing model. Meanwhile, standalone online marketplaces, like eBay and Etsy, have to revamp their entire business or exit the industry entirely, which further reduces competition for online retailing. The net competitive effects, then, are that consumers will pay higher prices, lose their ability to find long-tail items, and incur higher search costs to do so, while existing market leaders will consolidate their dominant positions, and hundreds of thousands of people will lose their jobs.

In contrast, who wins in this situation? The only winners are trademark owners, some of whom hate online marketplaces because they are tired of seeing their goods leak out of official distribution channels into more price-discounted online marketplaces, because they hate competing against used items of the goods they sell, and because some counterfeiting does take place there (as it does in the offline world too). To address those concerns, they are willing to burn down the entire online marketplace industry. What I can’t understand is why any members of Congress would be so willing to give trademark owners their wishlist when the results would be so disadvantageous for their constituents. The trademark owner lobby is strong, but our governance systems should be strong enough to resist terrible and selfish legislation like this.

Reposted with permission from Eric Goldman's Technology & Marketing Law Blog

25 Comments | Leave a Comment..

Posted on Techdirt - 14 April 2021 @ 10:48am

Deconstructing Justice Thomas' Pro-Censorship Statement

from the not-how-any-of-this-works dept

Last week, we had a post about Supreme Court Justice Clarence Thomas' very weird statement in a concurrence on mooting an unrelated case, in which he seemed to attack free speech and Section 230. Law professor Eric Goldman has written up an incredibly thorough response to Thomas' statement that we thought the Techdirt community might appreciate, and so we're reposting in here.

Last week, the Supreme Court vacated the Second Circuit’s Knight v. Trump ruling. The Second Circuit held that Trump violated the First Amendment when he blocked other Twitter users from engaging with his @realdonaldtrump account. Other courts are holding that government officials can’t block social media users from their official accounts, but they can freely block from personal or campaign accounts. Vacating the Second Circuit opinion probably won’t materially change that caselaw.

That outcome was overshadowed by a concurring statement from Justice Thomas wherein he again embraced censorship. I blogged a similar statement from Justice Thomas from the October 2020 cert denial of Enigma v. Malwarebytes. That time, Justice Thomas criticized Section 230 by addressing topics he wasn’t briefed on and clearly did not understand. This time, his statement is even more unhinged and disconnected from the case at issue. It’s clear Justice Thomas feels free to publish whatever thoughts are on his mind. This is what bloggers do. I think he, and all of us, would benefit if he moved his musings to a personal blog, instead of misusing our tax dollars to issue official government statements.

Justice Thomas’ statement ends (emphasis added):

As Twitter made clear, the right to cut off speech lies most powerfully in the hands of private digital platforms. The extent to which that power matters for purposes of the First Amendment and the extent to which that power could lawfully be modified raise interesting and important questions. This petition, unfortunately, affords us no opportunity to confront them.

So Justice Thomas acknowledges he wasn’t briefed on any of the interesting topics he wanted to discuss. He’s just making stuff up. This isn’t what Supreme Court justices do, or should do. I’m a little surprised that his colleagues haven’t publicly rebuked him for writing free-association statements. Such statements hurt the court’s credibility and abuse the privilege afforded Supreme Court justices.

Justice Thomas starts with an apparent contradiction he positions as a gotcha. The Second Circuit said that Trump created a public forum on Twitter, so Justice Thomas wonders how that could be when Twitter could unilaterally shut down that public forum. He says public forums are “government-controlled spaces,” but any “control Mr. Trump exercised over the account greatly paled in comparison to Twitter’s authority.” Still, Justice Thomas himself acknowledges that if the government rents private real property and uses it to create a public forum, it’s still a public forum even when a private landlord has the unilateral right to terminate the lease and evict the government. So….where’s the gotcha?

Having failed to define that problem, Justice Thomas manufactures a strawman. He says: “If part of the problem is private, concentrated control over online content and platforms available to the public, then part of the solution may be found in doctrines that limit the right of a private company to exclude.” Notice the conditional grammar to assume a problem without proving it. This is the foundation for a discussion about hypothetical solutions to hypothesized problems.

The two doctrines that “limit the right of a private company to exclude” are common carriage and public accommodations. That leads to this bone-chilling declaration:

Internet platforms of course have their own First Amendment interests, but regulations that might affect speech are valid if they would have been permissible at the time of the founding. See United States v. Stevens, 559 U. S. 460, 468 (2010). The long history in this country and in England of restricting the exclusion right of common carriers and places of public accommodation may save similar regulations today from triggering heightened scrutiny—especially where a restriction would not prohibit the company from speaking or force the company to endorse the speech. See Turner Broadcasting System, Inc. v. FCC, 512 U. S. 622, 684 (1994) (O’Connor, J., concurring in part and dissenting in part); PruneYard Shopping Center v. Robins, 447 U. S. 74, 88 (1980). There is a fair argument that some digital platforms are sufficiently akin to common carriers or places of accommodation to be regulated in this manner.

[Freeze frame and record scratch...] What did he just say?

First, notice how far Justice Thomas has strayed from the case before him. Somehow he’s talking about common carriage and public accommodations when neither doctrine had anything to do with Trump’s management of his Twitter account.

Second, did Justice Thomas just favorably cite Pruneyard? Most “conservatives” view Pruneyard skeptically because of its dramatic incursion into private property ownership. It’s also on the wane as precedent. Courts have been reluctant to extend it to new facts. The Pruneyard decision may be a low-water mark for private property ownership rights, not the foundation of expanded censorship. (There is also the standard Internet exceptionalism problem with applying an offline analogy like physical-space shopping malls to online media venues).

Third, he is about to make a “fair argument” that “some digital platforms are sufficiently akin to common carriers or places of accommodation.” OK, but are there any counterarguments to that “fair” argument? Normally an opposing litigant would be aggressively telling its side of the story, and other Supreme Court justices would be pointing out the weaknesses in Justice Thomas’ “fair” arguments. Without these tempering forces, Justice Thomas is engaging in personal advocacy, not judicial analysis.

Regarding common carriers, Justice Thomas claims:

In many ways, digital platforms that hold themselves out to the public resemble traditional common carriers. Though digital instead of physical, they are at bottom communications networks, and they “carry” information from one user to another…unlike newspapers, digital platforms hold themselves out as organizations that focus on distributing the speech of the broader public.

It should not matter how an editorial publication sources the content it publishes. I remember Zagat, which tried to faithfully mirror the opinions of ordinary restaurant consumers. Did “distributing the speech of the broader public” make Zagat a common carrier? Of course not, because Zagat layered substantial editorial value on top of the consumer comments. But so does Twitter, which enforces its house rules and performs many crucial curatorial functions. Justice Thomas ignores those value-added editorial functions.

Justice Thomas then links common carriage to network effects:

The analogy to common carriers is even clearer for digital platforms that have dominant market share. Similar to utilities, today’s dominant digital platforms derive much of their value from network size….The Facebook suite of apps is valuable largely because 3 billion people use it. Google search—at 90% of the market share—is valuable relative to other search engines because more people use it, creating data that Google’s algorithm uses to refine and improve search results. These network effects entrench these companies. Ordinarily, the astronomical profit margins of these platforms—last year, Google brought in $182.5 billion total, $40.3 billion in net income—would induce new entrants into the market. That these companies have no comparable competitors highlights that the industries may have substantial barriers to entry….

It changes nothing that these platforms are not the sole means for distributing speech or information. A person always could choose to avoid the toll bridge or train and instead swim the Charles River or hike the Oregon Trail. But in assessing whether a company exercises substantial market power, what matters is whether the alternatives are comparable. For many of today’s digital platforms, nothing is.

The companies Justice Thomas disparages would hotly contest his assessment. But they weren’t in his courtroom to explain themselves.

More generally, normally common carriage redresses natural monopolies, where it would be socially wasteful to build duplicative infrastructure. Assuming Facebook and Google in fact benefit from network effects, they still lack that key attribute of natural monopolists. In particular, competitors can and will successfully compete by providing non-identical orthogonal solutions.

Justice Thomas continues smearing non-litigants:

Much like with a communications utility, this concentration gives some digital platforms enormous control over speech. When a user does not already know exactly where to find something on the Internet—and users rarely do—Google is the gatekeeper between that user and the speech of others 90% of the time. It can suppress content by deindexing or downlisting a search result or by steering users away from certain content by manually altering autocomplete results. Grind, Schechner, McMillan, & West, How Google Interferes With Its Search Algorithms and Changes Your Results, Wall Street Journal, Nov. 15, 2019. Facebook and Twitter can greatly narrow a person’s information flow through similar means. And, as the distributor of the clear majority of e-books and about half of all physical books, Amazon can impose cataclysmic consequences on authors by, among other things, blocking a listing.

Is Justice Thomas suggesting all of these services–including Amazon’s book retailing–should be treated like common carriers? Where does that stop?

Also, media industry consolidation is ubiquitous in every media niche. For example, there are 3 major record labels, and Disney has eaten a huge chunk of the movie business. Does that make them common carriers? In the 1970s and 1980s, there was a single daily newspaper in each metro area. Should they have been deemed common carriers because of that? Recall Florida tried to do that in Miami Herald v. Tornillo (though it didn’t use the term “common carrier”). The Supreme Court held that the Miami Herald newspaper’s local market dominance did not reduce the newspaper’s constitutional protection.

With respect to public accommodations, Justice Thomas says “a company ordinarily is a place of public accommodation if it provides ‘lodging, food, entertainment, or other services to the public . . . in general.’ Twitter and other digital platforms bear resemblance to that definition.” Every business will bear some “resemblance” to that definition because they offer goods or services to their customers, but not every business is a place of public accommodation. Justice Thomas closes the thought by saying “no party has identified any public accommodation restriction that applies here.” That’s because IT WASN’T RELEVANT TO THE CASE.

Justice Thomas cheerleads the #MAGA legislators around the country working on censorial bills:

The similarities between some digital platforms and common carriers or places of public accommodation may give legislators strong arguments for similarly regulating digital platforms. “[I]t stands to reason that if Congress may demand that telephone companies operate as common carriers, it can ask the same of” digital platforms. Turner, 512 U. S., at 684 (opinion of O’Connor, J.). That is especially true because the space constraints on digital platforms are practically nonexistent (unlike on cable companies), so a regulation restricting a digital platform’s right to exclude might not appreciably impede the platform from speaking.

Justice Thomas somehow overlooked Reno v. ACLU (1997), which came out after Turner and Denver Area. The Supreme Court said that, unlike broadcasting and telecom, there was no basis for qualifying the First Amendment scrutiny applied to Internet content regulations. This is 100% responsive to his invocation of O’Connor’s language from Turner.

Justice Thomas then says “plaintiffs might have colorable claims against a digital platform if it took adverse action against them in response to government threats.” Not this again. It’s a true statement with respect to “government threats,” but general censorial exhortations by government officials aren’t “threats.” In a footnote, he adds:

Threats directed at digital platforms can be especially problematic in the light of 47 U. S. C. §230, which some courts have misconstrued to give digital platforms immunity for bad-faith removal of third-party content. Malwarebytes, Inc. v. Enigma Software Group USA, LLC, 592 U. S. ___, ___–___ (2020) (THOMAS, J., statement respecting denial of certiorari) (slip op., at 7–8). This immunity eliminates the biggest deterrent—a private lawsuit—against caving to an unconstitutional government threat.

Wait, who is the villain in that story? My vote: The government making unconstitutional threats. Section 230 doesn’t prevent lawsuits directly against the government for issuing these threats. Nevertheless, Justice Thomas apparently thinks that Internet services, receiving unconstitutional demands from government officials, should be sued by individual users for honoring those demands. Yet, an Internet service’s content removal in response to a government threat usually would be considered a “good faith” removal and thus satisfy the statutory requirements of Section 230(c)(2), so I don’t understand why Justice Thomas thinks his Enigma statement is relevant. And if Section 230 didn’t protect the Internet service’s removal, is Justice Thomas saying that the Internet services should be compelled to carry potentially illegal content even if the government executes its threat? Here’s a better idea: we should all work together to stop the government from issuing unconstitutional threats. And the first government threat I think we should stop? I nominate Justice Thomas’ threat to impose must-carry obligations.

Justice Thomas, citing Prof. Volokh, speculates that maybe Section 230 is itself unconstitutional:

some commentators have suggested that immunity provisions like §230 could potentially violate the First Amendment to the extent those provisions pre-empt state laws that protect speech from private censorship

As I’ve said before, the phrase “private censorship” is an oxymoron. Only governments censor. Private entities exercise editorial control.

More generally, I do not see how Section 230(c)(1) is unconstitutional. It’s a speech-enhancing statute that supplements the First Amendment. Section 230(c)(2) is more colorable because it does make distinctions between different content categories. However, so long as courts read the “otherwise objectionable” exclusion broadly, that phrase basically applies to all content equally. Note that various Section 230(c)(2) reforms propose to remove or modify the “otherwise objectionable” language, and those changes could create a constitutional problem where none currently exists.

Justice Thomas says the threats he’s talking about have nothing to do with the case at hand:

But no threat is alleged here…no party has sued Twitter. The question facing the courts below involved only whether a government actor violated the First Amendment by blocking another Twitter user.

I agree. So why is Justice Thomas discussing any of this?

Justice Thomas’ statement concludes:

The Second Circuit feared that then-President Trump cut off speech by using the features that Twitter made available to him. But if the aim is to ensure that speech is not smothered, then the more glaring concern must perforce be the dominant digital platforms themselves. As Twitter made clear, the right to cut off speech lies most powerfully in the hands of private digital platforms.

I strongly disagree about the MOST “glaring concern” here. Twitter lacks the power to order drone killings, separate parents from their children at the border, put a knee on the neck of a suspect for 9 minutes, incarcerate people, impose taxes, garnish people’s wages, or engage in the thousands of other ways that governments can deprive people of their assets, liberty, or life. Compared to the government’s vast power to squelch speech, the power of the “dominant digital platforms” seems puny. Justice Thomas betrays his extraordinary degree of privilege. Due to that privilege, he doesn’t recognize how the truly glaring concern is that the government, fueled by his words, will use its “dominance” to “smother” far more speech than any Internet service ever could.


I hope Justice Thomas’ colleagues do not share his views and this statement is just idle musings. But even if the statement doesn’t lead to changes at the Supreme Court, it will nevertheless contribute to three unfortunate dynamics.

First, plaintiffs will improperly cite the statement as if it is binding law (which they did with his prior statement: 1, 2). They will especially like the discussion about government threats.

Second, plaintiffs will appeal more censorial cases to the Supreme Court, knowing that Justice Thomas is a reliable vote to grant the cert petition and vote in their favor.

Third, state legislators will view this opinion as permission to pursue unconstitutional must-carry obligations. There are so many proposals percolating in the state legislatures right now, and odds are good that at least one will get enacted and the battle will shift to the court challenges of those laws. The future of the Internet rests on those coming court battles, and I feel less secure about the Internet’s fate knowing that Justice Thomas is one of the final 9 votes.

Finally, remember that Trump’s Twitter account was government speech. The thrust of Justice Thomas’ statement would require Twitter to carry government speech it doesn’t want to carry. That isn’t garden-variety censorship. Justice Thomas seemingly wants private media operations to become government mouthpieces. Forcing media outlets to distribute government propaganda is a hallmark of repressive and autocratic countries. I don’t know what it means to be a “conservative,” but I know it shouldn’t include that.

BONUS: Justice Thomas isn’t trying to hide his antipathy towards Google. See this passage from his dissent in Google v. Oracle, No. 18–956 (U.S. Sup. Ct. April 5, 2021):

If the majority is going to speculate about what Oracle might do, it at least should consider what Google has done. The majority expresses concern that Oracle might abuse its copyright protection (on outdated Android versions) and “‘attempt to monopolize the market.’” Ante, at 34–35. But it is Google that recently was fined a record $5 billion for abusing Android to violate antitrust laws. Case AT.40099, Google Android, July 18, 2018 (Eur. Comm’n-Competition); European Comm’n Press Release, Commission Fines Google €4.34 Billion for Illegal Practices Regarding Android Mobile Devices to Strengthen Dominance of Google’s Search Engine, July 18, 2018. Google controls the most widely used mobile operating system in the world. And if companies may now freely copy libraries of declaring code whenever it is more convenient than writing their own, others will likely hesitate to spend the resources Oracle did to create intuitive, well-organized libraries that attract programmers and could compete with Android. If the majority is worried about monopolization, it ought to consider whether Google is the greater threat.

Originally posted to Eric Goldman's Technology & Marketing Law Blog.

48 Comments | Leave a Comment..

Posted on Techdirt - 2 December 2020 @ 7:00am

New Ebook On Zeran v AOL, The Most Important Section 230 Case

from the read-it,-understand-it dept

Section 230 has become a mainstream discussion topic, but unfortunately many discussants don’t actually understand it well (or at all). To address this knowledge gap, co-editors Profs. Eric Goldman (Santa Clara Law) and Jeff Kosseff (U.S. Naval Academy) have released an ebook, called “Zeran v. America Online,” addressing many aspects of Section 230. You can download the ebook for free at:

Zeran v. AOL is the most important Section 230 case of all time. The Zeran case was the first federal appellate decision interpreting Section 230, and its breathtakingly broad sweep turbocharged the rise of Web 2.0–with all of its strengths and weaknesses.

In recognition of Zeran’s importance, in 2017, Profs. Goldman and Kosseff helped assemble an essay package to honor the case’s 20th anniversary. They gathered two dozen essays from some of the most knowledgeable Section 230 experts. The essays address the case’s history and policy implications. Initially, published the essay package but then unexpectedly paywalled the essays after 6 months.

The new Zeran v. America Online ebook restores the 2017 essay package into a new and easy-to-read format. Together, they are a great entry point into the debates about Section 230, including how we got here and what’s at stake.

To supplement the essay package, the ebook compiles an archive of the key documents in the Zeran v. AOL litigation (with bonus coverage of Zeran’s case against radio station KRXO). Many of these materials have not previously been publicly available in electronic format. The case archives should be of interest both to historians and students of precedent-setting litigation tactics.

Section 230 will likely remain hotly debated, but the debates won’t be productive until we develop a shared understanding of what the law says and why. Ideally, this ebook will advance those goals.

Read More | 2 Comments | Leave a Comment..

Posted on Techdirt - 16 October 2019 @ 6:38am

Top Myths About Content Moderation

from the so-many-myths-to-debunk dept

How Internet companies decide which user-submitted content to keep and which to remove—a process called “content moderation”—is getting lots of attention lately, for good reason. Under-moderation can lead to major social problems, like foreign agents manipulating our elections. Over-moderation can suppress socially beneficial content, like negative but true reviews by consumers.

Due to these high stakes, regulators across the globe increasingly seek to tell Internet companies how to moderate content. European regulators are requiring Internet services to remove extremist content within an hour and to install upload filters to prospectively block copyright infringement; and U.S. legislators have proposed to ban Internet services from moderating content at all.

Unfortunately, many of these regulatory efforts are predicated on myths about content moderation, such as:

Myth: Content moderation can be done perfectly.

Reality: Regulators routinely assume Internet services can remove all bad content without suppressing any good content. Unfortunately, they can’t. First, mistakes occur when the service lacks key contextual information about the content—such as details about the author’s identity, other online and offline activities, and cultural references. Second, any line-drawing exercise creates mistake-prone border cases because users routinely submit “edgy” content. Third, a high-volume service will make many mistakes, even if it’s highly accurate—1 billion submissions a day at 99.9% accuracy still yields a million mistakes a day.

Myth: Bad content is easy to find and remove.

Reality: Regulators often assume every item of bad content has an impossible-to-miss flashing neon sign saying “REMOVE THIS CONTENT,” but that’s rare. Content is often obviously bad only in hindsight or with context unavailable to the service. Regulators’ cherry-picked anecdotes don’t prove otherwise.

Myth: Technologists just need to “nerd harder.”

Reality: Filtering and artificial intelligence play important roles in content moderation. However, technology alone cannot magically solve the problem. “Edgy” and contextless content vexes the machines, too.

Myth: Internet services should hire more humans to review content.

Reality: Humans have biases and make mistakes too, so adding human reviewers won’t lead to perfection. Furthermore, human reviewers sometimes experience an unrelenting onslaught of horrible content to protect the rest of us.

Myth: Internet companies have no incentive to moderate content.

Reality: In 1996, Congress passed 47 U.S.C. 230, which says Internet services generally aren’t liable for third-party content. Due to this legal protection, critics often assume Internet services won’t invest in content moderation; and some companies have stoked that perception by publicly positioning themselves as “neutral” technology platforms. Yet, virtually every Internet service moderates content, and major services like Facebook and YouTube employ many thousands of content reviewers. Why? The services have their own reputation to manage, and they care about how content can affect their users (e.g., Pinterest combats content that promotes eating disorders). Furthermore, advertisers won’t let their ads appear on bad content, which provides additional financial incentives to moderate.

Myth: Content moderation, if done right, will make everyone happy.

Reality: By definition, content moderation is a zero-sum game. Someone gets their desired outcome, and someone else doesn’t—and those folks won’t be happy with the result.

Myth: There is a one-size-fits-all approach to content moderation.

Reality: Internet services cater to diverse audiences that have different moderation needs. For example, an online crowdsourced encyclopedia like Wikipedia, an open-source software repository like GitHub, and a payment service for content publishers like Patreon all solve different problems for their communities. These services shouldn’t have identical content moderation rules.

Myth: Imposing content moderation requirements will stick it to Google and Facebook.

Reality: Google and Facebook have enough money to handle virtually any requirement imposed by regulators. Startup enterprises do not. Increased content moderation burdens are more likely to block new entrants than to punish Google and Facebook.

Myth: Poor content moderation causes anti-social behavior.

Reality: Poorly executed content moderation can accelerate bad behavior, but often the Internet simply mirrors existing anti-social behavior or tendencies. Better content moderation can’t fix problems that are endemic in the human condition.

Regulators are right to identify content moderation as a critically important topic. However, until regulators overcome these myths, regulatory interventions will cause more problems than they solve.

Reposted from Eric Goldman's Technology & Marketing Law Blog.

94 Comments | Leave a Comment..

Posted on Techdirt - 8 November 2018 @ 2:52pm

CDA 230 Doesn't Support Habeus Petition by 'Revenge Pornographer'

from the cda-230-still-not-absolute dept

As you may recall, Kevin Bollaert ran UGotPosted, which published third-party submitted nonconsensual pornography, and, which offered depicted individuals a "pay-to-remove" option. Bollaert appeared multiple times in my inventory of nonconsensual pornography enforcement actions. Bollaert's conduct was disgusting, and I have zero sympathy for him. Nevertheless, I also didn't love the path prosecutors took to bust him. The lower court convicted him of 24 counts of identity theft and 7 counts of extortion and sentenced him to 8 years in jail and 10 years of supervised release. Pay-to-remove sites are not inherently extortive, and identity theft crimes often overreach to cover distantly related activities.

Worse, the appeals court affirmed the convictions despite a significant Section 230 defense. The opinion contorted Section 230 law, relying on outmoded legal theories from Fortunately, I haven't seen many citations to the appellate court's misinterpretation of Section 230, so the doctrinal damage to Section 230 hasn't spread too much (yet). However, that still leaves open whether Bollaert's conviction was correct.

Bollaert raised that issue by filing a habeus corpus petition in federal court. Such petitions are commonly filed and almost never granted, so Bollaert's petition had minimal odds of success as a matter of math. Not surprisingly, his petition fails.

The district court says that Section 230's application to Bollaert's circumstance does not meet the rigorous standard of "clearly established federal law":

In this case, the Supreme Court has never recognized that the CDA applies in state criminal actions. The Supreme Court has never indicated circumstances that would qualify a state criminal defendant for CDA immunity. Absence of applicable Supreme Court precedent defeats the contention that Petitioner is entitled to CDA immunity under clearly established federal law...

federal circuits have not applied CDA immunity in state criminal actions or indicated circumstances that would qualify a state criminal defendant for CDA immunity. Petitioner cannot satisfy § 2254(d)(1) with district court opinions applying CDA immunity in state criminal actions.

I've routinely blogged about the application of Section 230 to state criminal prosecutions, and I even wrote a lengthy discourse on why that was a good thing. Still, I can't think of any federal appellate courts that have reached this conclusion, so perhaps the court's factual claim about the jurisprudential absence is correct.

The court adds that even if Section 230 qualified as "clearly established federal law," the appellate court ruling didn't necessarily contravene that law:

the California Court of Appeal performed an exhaustive and comprehensive analysis of the applicable circuit court decisions before concluding Petitioner is an information content provider under Roommates. The state court reasonably interpreted Roommates and Jones, and reasonably concluded that Petitioner "developed, at least in part, the offensive content on his Web site by requiring users to input private and personal information as a condition of posting the victims' pictures, making him an information content provider within the meaning of the CDA."

This passage reinforces the deficiencies of the appellate court's Section 230 discussion. "[R]equiring users to input private and personal information as a condition of posting the victims' pictures" is not the encouragement of illegal content, as referenced by, as that information isn't actually illegal; and the Jones case rejected an "encouragement" exclusion to Section 230 while ruling for the defense. Do those deficiencies support the extraordinary relief of habeus corpus? Apparently not.

Reposted from Eric Goldman's Technology & Marketing Law Blog

Read More | 31 Comments | Leave a Comment..

Posted on Techdirt - 29 January 2018 @ 1:23pm

It's Time to Talk About Internet Companies' Content Moderation Operations

from the transparency dept

As discussed in this post below, on February 2nd, Santa Clara University is hosting a gathering of tech platform companies to discuss how they actually handle content moderation questions. Many of the participants have written short essays about the questions that will be discussed at this event -- and over the next few weeks we'll be publishing many of those essays. This first one comes from Professor Eric Goldman, who put together the conference, explaining the rationale behind the event and this series of essays.

Many user-generated content (UGC) services aspire to build scalable businesses where usage and revenues grow without increasing headcount. Even with advances in automated filtering and artificial intelligence, this goal is not realistic. Large UGC databases require substantial human intervention to moderate anti-social and otherwise unwanted content and activities. Despite the often-misguided assumptions by policymakers, problematic content usually does not have flashing neon signs saying "FILTER ME!" Instead, humans must find and remove that content—especially with borderline cases, where machines can't make sufficiently nuanced judgments.

At the largest UGC services, the number of people working on content moderation is eye-popping. By 2018, YouTube will have 10,000 people on its "trust & safety teams." Facebook's "safety and security team" will grow to 20,000 people in 2018.

Who are these people? What exactly do they do? How are they trained? Who sets the policies about what content the service considers acceptable?

We have surprisingly few answers to these questions. Occasionally, companies have discussed these topics in closed-door events, but very little of this information has been made public.

This silence is unfortunate. A UGC service's decision to publish or remove content can have substantial implications for individuals and the community, yet we lack the information to understand how those decisions are made and by whom. Furthermore, the silence has inhibited the development of industry-wide "best practices." UGC services can learn a lot from each other—if they start sharing information publicly.

On Friday, a conference called "Content Moderation and Removal at Scale" will take place at Santa Clara University. (The conference is sold out, but we will post recordings of the proceedings, and we hope to make a live-stream available). Ten UGC services will present "facts and figures" about their content moderation operations, and five panels will discuss cutting-edge content moderation issues. For some services, this conference will be the first time they've publicly revealed details about their content moderation operations. Ideally, the conference will end the industry's norm of silence.

In anticipation of the conference, we assembled ten essays from conference speakers discussing various aspects of content moderation. These essays provide a sample of the conversation we anticipate at the conference. Expect to hear a lot more about content moderation operational issues in the coming months and years.

Eric Goldman is a Professor of Law, and Co-Director of the High Tech Law Institute, at Santa Clara University School of Law. He has researched and taught Internet Law for over 20 years, and he blogs on the topic at the Technology & Marketing Law Blog.

66 Comments | Leave a Comment..

Posted on Techdirt - 12 April 2017 @ 2:58pm

Texas Supreme Court Is Skeptical About Wikipedia As A Dictionary

from the the-things-courts-debate dept

This is an interesting opinion from the Texas Supreme Court on citing Wikipedia as a dictionary. The underlying case involves an article in D Magazine titled "The Park Cities Welfare Queen." The article purports to show that the plaintiff, Rosenthal, "has figured out how to get food stamps while living in the lap of luxury." After publication, evidence emerged that the plaintiff had not committed welfare fraud. She sued the magazine for defamation.

The appeals court denied the magazine's anti-SLAPP motion in part because it held the term "Welfare Queen," as informed by the Wikipedia entry, could be defamatory. The Texas Supreme Court affirms the anti-SLAPP denial, but it also criticizes the appeals court for not sufficiently examining the entire article's gist. Along the way, the court opines on the credibility and validity of Wikipedia as a dictionary. TL;DR = the Supreme Court says don't treat Wikipedia like a dictionary.

Apologies for the block quoting, but here's the detail:

Wikipedia is a self-described "online open-content collaborative encyclopedia." Wikipedia: General Disclaimer, (last visited Mar. 13, 2017). This means that, except in certain cases to prevent disruption or vandalism, anyone can write and make changes to Wikipedia pages. Wikipedia: About, (last visited Mar. 13, 2017). Volunteer editors can submit content as registered members or anonymously. Id. Each time an editor modifies content, the editor's identity or IP address and a summary of the modification, including a time stamp, become available on the article's "history" tab. Jason C. Miller & Hannah B. Murray, Wikipedia in Court: When and How Citing Wikipedia and Other Consensus Websites Is Appropriate, 84 ST. JOHN'S L. REV. 633, 637 (2010). Wikipedia is one of the largest reference websites in the world, with over "70,000 active contributors working on more than 41,000,000 articles in 294 languages." Wikipedia: About, supra.

References to Wikipedia in judicial opinions began in 2004 and have increased each year, although such references are still included in only a small percentage of opinions. Jodi L. Wilson, Proceed with Extreme Caution: Citation to Wikipedia in Light of Contributor Demographics and Content Policies, 16 VAND. J. ENT. & TECH. L. 857, 868 (2014). These cites often relate to nondispositive matters or are included in string citations. But, some courts "have taken judicial notice of Wikipedia content, based their reasoning on Wikipedia entries, and decided dispositive motions on the basis of Wikipedia content." Lee F. Peoples, The Citation of Wikipedia in Judicial Opinions, 12 YALE J. L. & TECH. 1, 3 (2009–2010). While there has been extensive research on Wikipedia's accuracy, "the results are mixed—some studies show it is just as good as the experts, [while] others show Wikipedia is not accurate at all." Michael Blanding, Wikipedia or Encyclopædia Britannica: Which Has More Bias?, FORBES (Jan. 20, 2015),

Any court reliance on Wikipedia may understandably raise concerns because of "the impermanence of Wikipedia content, which can be edited by anyone at any time, and the dubious quality of the information found on Wikipedia." Peoples, supra at 3. Cass Sunstein, legal scholar and professor at Harvard Law School, also warns that judges' use of Wikipedia "might introduce opportunistic editing." Noam Cohen, Courts Turn to Wikipedia, but Selectively, N.Y. TIMES (Jan. 29, 2007), 29wikipedia.html. The Fifth Circuit has similarly warned against using Wikipedia in judicial opinions, agreeing "with those courts that have found Wikipedia to be an unreliable source of information" and advising "against any improper reliance on it or similarly unreliable internet sources in the future." Bing Shun Li v. Holder, 400 F. App'x 854, 857 (5th Cir. 2010); accord Badasa v. Mukasey, 540 F.3d 909, 910–11 (8th Cir. 2008).

For others in the legal community, however, Wikipedia is a valuable resource. Judge Richard Posner has said that "Wikipedia is a terrific resource … because it [is] so convenient, it often has been updated recently and is very accurate." Cohen, supra. However, Judge Posner also noted that it "wouldn't be right to use it in a critical issue." Id. Other scholars agree that Wikipedia is most appropriate for "soft facts," when courts want to provide context to help make their opinions more readable. Id. Moreover, because Wikipedia is constantly updated, some argue that it can be "a good source for definitions of new slang terms, for popular culture references, and for jargon and lingo including computer and technology terms." Peoples, supra at 31. They also argue that open-source tools like Wikipedia may be useful when courts are trying to determine public perception or community norms. Id. at 32. This usefulness is lessened, however, by the recognition that Wikipedia contributors do not necessarily represent a cross-section of society, as research has shown that they are overwhelmingly male, under forty years old, and living outside of the United States. Wilson, supra at 885–89.

Given the arguments both for and against reliance on Wikipedia, as well as the variety of ways in which the source may be utilized, a bright-line rule is untenable. Of the many concerns expressed about Wikipedia use, lack of reliability is paramount and may often preclude its use as a source of authority in opinions. At the least, we find it unlikely Wikipedia could suffice as the sole source of authority on an issue of any significance to a case. That said, Wikipedia can often be useful as a starting point for research purposes. See Peoples, supra at 28 ("Selectively using Wikipedia for … minor points in an opinion is an economical use of judges' and law clerks' time."). In this case, for example, the cited Wikipedia page itself cited past newspaper and magazine articles that had used the term "welfare queen" in various contexts and could help shed light on how a reasonable person could construe the term.

However, the court of appeals utilized Wikipedia as its primary source to ascribe a specific, narrow definition to a single term that the court found significantly influenced the article's gist. Essentially, the court used the Wikipedia definition as the lynchpin of its analysis on a critical issue. As a result, the court narrowly read the term "welfare queen" to necessarily implicate fraudulent or illegal conduct, while other sources connote a broader common meaning. See, e.g., Oxford Living Dictionaries, (last visited Mar. 13, 2017) (broadly defining "welfare queen" as a "woman perceived to be living in luxury on benefits obtained by exploiting or defrauding the welfare system"); YourDictionary, (last visited Mar. 13, 2017) (broadly defining "welfare queen" as a "woman collecting welfare, seen as doing so out of laziness, rather than genuine need"). In addition, and independent of the Wikipedia concerns, the court of appeals' overwhelming emphasis on a single term in determining the article's gist departed from our jurisprudential mandate to evaluate the publication as a whole rather than focus on individual statements.

A concurring opinion by Justice Guzman amplifies the concerns (FNs omitted):

Wikipedia has many strengths and benefits, but reliance on unverified, crowd-generated information to support judicial rulings is unwise. Mass-edited collaborative resources, like Wikipedia, are malleable by design, raising serious concerns about the accuracy and completeness of the information, the expertise and credentials of the contributors, and the potential for manipulation and bias. In an age when news about "fake news" has become commonplace, long-standing concerns about the validity of information obtained from "consensus websites" like Wikipedia are not merely the antiquated musings of luddites. To the contrary, as current events punctuate with clarity, courts must remain vigilant in guarding against undue reliance on sources of dubious reliability. A collaborative encyclopedia that may be anonymously and continuously edited undoubtedly fits the bill.

Legal commentators may debate whether and to what extent courts could properly rely on online sources like Wikipedia, but the most damning indictment of Wikipedia's authoritative force comes directly from Wikipedia:

  • "Please be advised that nothing found here has necessarily been reviewed by people with the expertise required to provide you with complete, accurate or reliable information."
  • "Wikipedia cannot guarantee the validity of the information found here."
  • "Wikipedia is not uniformly peer reviewed."
  • "[A]ll information read here is without any implied warranty of fitness for any purpose or use whatsoever."
  • "Even articles that have been vetted by informal peer review or featured article processes may later have been edited inappropriately, just before you view them."
  • Indeed, "Wikipedia's radical openness means that any given article may be, at any given moment, in a bad state: for example, it could be in the middle of a large edit or it could have been recently vandalized." Even if expeditiously remediated, transient errors are not always obvious to the casual reader. As Wikipedia states more pointedly, "Wikipedia is a wiki, which means that anyone in the world can edit an article, deleting accurate information or adding false information, which the reader may not recognize. Thus, you probably shouldn't be citing Wikipedia."

    Apart from these candid self-assessments, which no doubt apply with equal force to other online sources and encyclopedias, a more pernicious evil lurks—"opportunistic editing." Because "[a]nyone with Internet access can write and make changes to Wikipedia articles" and "can contribute anonymously, [or] under a pseudonym," reliance on Wikipedia as an authoritative source for judicial decision-making incentivizes self-interested manipulation. Case in point: a Utah court of appeals recently described how the Wikipedia definition of "jet ski" provided "stronger support" for one of the parties in a subsequent appeal than it had when considered by the court in the parties' previous appeal. The court observed the difficulty of discerning whether the change was instigated by the court's prior opinion, perhaps "at the instance of someone with a stake in the debate."

    Still, some have argued Wikipedia is "a good source for definitions of new slang terms, for popular culture references, and for jargon and lingo including computer and technology terms." Perhaps, but not necessarily. While Wikipedia's "openly editable" model may be well suited to capturing nuances and subtle shifts in linguistic meaning, there is no assurance that any particular definition actually represents the commonly understood meaning of a term that may be central to a legal inquiry. In truth, Wikipedia's own policies disclaim the notion: "Wikipedia is not a dictionary, phrasebook, or a slang, jargon or usage guide." Whatever merit there may be to crowdsourcing the English language, Wikipedia simply lacks the necessary safeguards to prevent abuse and assure the level of certainty and validity typically required to sustain a judgment in a legal proceeding.

    Take, for example, the Wikipedia entry for "welfare queen," which was first created in November 2006 by the user Chalyres. Since the entry was first drafted, 239 edits have been made by 146 users. But there is no reliable way to determine whether these edits (1) deleted or added accurate information, (2) deleted or added false or biased information, (3) were made by individuals with expertise on the term's usage, or (4) were made by individuals actually representative of the community.

    As a court, one of our "chief functions" is "to act as an animated and authoritative dictionary." In that vein, we are routinely called upon to determine the common meaning of words and phrases in contracts, statutes, and other legal documents. Though we often consult dictionaries in discharging our duty, rarely, if ever, is one source alone sufficient to fulfill the task. To that end, I acknowledge that Wikipedia may be useful as a "starting point for serious research," but it must never be considered "an endpoint," at least in judicial proceedings.

    Wikipedia's valuable role in today's technological society cannot be denied. Our society benefits from the fast, free, and easily-accessible information it provides. A wealth of information is now available at the touch of a few key strokes, and a community of Wikipedia editors serves to increase the accuracy and truth of that information, promoting the public good through those efforts. However, in my view, Wikipedia properly serves the judiciary only as a compendium—a source for sources—and not as authority for any disputed, dispositive, or legally consequential matter.

    To punctuate her skepticism, Justice Guzman's concurrence displays this screenshot:

    In a footnote, you can almost hear a sneer as she characterizes the screenshot as "Screenshot of unsaved edits to Welfare Queen." NB: Wikipedia is trivially easy to edit, but getting those edits to stick is an entirely different matter.

    My Thoughts

    It makes sense not to treat Wikipedia as the authoritative citation source. However, I would make the same declaration about many sources, crowd-sourced or not. Often, a range of sources is required to establish a "fact."

    We especially see the trickiness of treating a single dictionary as an authoritative source, because there are often subtle but crucial differences in dictionaries' definitions of the same term. Indeed, Wikipedia self-acknowledges its limits as a dictionary. In contrast, sometimes Wikipedia is an OK citation for the zeitgeist about an issue, where the citation is for the ranges of issues rather than for the truth of any issue.

    I was a little surprised that the court didn't discuss the Urban Dictionary as an alternative to Wikipedia as a dictionary (it comes up only in a reference in a footnote in Justice Guzman's opinion). What I like about Urban Dictionary is that it doesn't purport to offer a single definition of any term. Instead, it lists a range of definitions ordered by crowd-sourced voting. In my experience, the Urban Dictionary often fills in the gaps in my "street lingo" much better than any other source, so long as I use it advisedly.

    I'm paying closer attention to courts' citations to online dictionaries based on my research for my Emojis and the Law paper. As bad as things are between Wikipedia and Urban Dictionary as online dictionaries, things are much worse with emojis because no credible dictionary is trying to provide definitive definitions of emojis. Eventually, as I'll argue in my paper, we'll need the equivalent of an Urban Dictionary for emojis to capture their disparate meanings across online subcommunities.

    Republished from Eric Goldman's Technology & Marketing Law Blog

    Read More | 16 Comments | Leave a Comment..

    Posted on Techdirt - 28 September 2016 @ 1:02pm

    Does The FTC Get To Ignore Section 230 Of The CDA?

    from the apparently dept

    I've often joked that the FTC and state AGs choose to live in a fantasy world where Section 230 doesn't exist. A new ruling from the Second Circuit has turned my joke on its ear, suggesting that my underlying fears -- of a Section 230-free zone for consumer protection agencies -- may have become our dystopian reality.

    The Opinion

    The case involves weight loss products, including colon cleanses, vended by LeanSpa. To generate more sales, LeanSpa hired LeadClick to act as an affiliate marketing manager. LeadClick coordinated promotion of LeanSpa's products with LeadClick's network of affiliates. Some affiliates promoted the products using fake news sites, with articles styled to look like legitimate news articles and consumer comments/testimonials that were fake. Apparently, all of this added up to big business. LeanSpa paid LeadClick $35-$45 each time a consumer signed up for LeanSpa's "free" trial (which was a negative billing option). LeadClick shared 80-90% of these sign-up fees with affiliates and kept the remainder for itself. In total, LeadClick billed LeanSpa $22M, of which LeanSpa paid only $12M. Still, LeanSpa turned into LeadClick's top customer, constituting 85% of its eAdvertising division's sales.

    The court summarizes the key facts about LeadClick's role in the fake new sites scheme:

    While LeadClick did not itself create fake news sites to advertise products…it (1) knew that fake news sites were common in the affiliate marketing industry and that some of its affiliates were using fake news sites, (2) approved of the use of these sites, and, (3) on occasion, provided affiliates with content to use on their fake news pages.

    The court also notes that LeadClick occasionally bought ads on legitimate news sites to promote fake news sites in its affiliate network.

    The FTC's Prima Facie Case

    The FTC alleged that LeadClick engaged in deceptive practices. LeadClick responded that it didn't do any deceptive practices itself; if anyone did, it was its affiliates. Extensively citing the Ninth Circuit's FTC v. Neovi ruling from 2010 (an unfairness case, not a deception case, but this panel ignores the difference) and a subsequent 11th Circuit case (FTC v. IAB Marketing Associates), the Second Circuit concludes that "a defendant may be held liable for engaging in deceptive practices or acts if, with knowledge of the deception, it either directly participates in a deceptive scheme or has the authority to control the deceptive content at issue."

    In the Neovi case, the defendant Qchex had an online check-creation tool that fraudsters used to create and send bogus checks. The court held that Qchex engaged in unfair practices when it printed and then delivered the bogus checks to recipients. But here, LeadClick never "delivered" anything. Indeed, LeadClick argued that the legal standard conflates direct liability with aiding/abetting liability. The Second Circuit disagreed, saying a defendant who "allows the deception to proceed" thus "engages, through its own actions, in a deceptive act or practice that causes harm to consumers."

    I'm not a philosopher, but to me, "allowing" a third party to commit misconduct is a bizarre and overly expansive way of defining *direct* liability. Once this court makes this doctrinal cheat, LeadClick didn't have a chance. Applying the legal standard to LeadClick:

    • knowledge. "LeadClick knew that (1) the use of false news pages was prevalent in affiliate marketing, and (2) its own affiliate marketers were using fake news sites to market LeanSpaʹs products."
    • "direct participation in the deceptive conduct." LeadClick satisfied this standard by "recruiting and paying affiliates who used fake news sites for generating traffic, managing those affiliates, suggesting substantive edits to fake news pages, and purchasing banner space for fake news sites on legitimate news sources."
    • "ability to control." LeadClick ran an affiliate network that included fake news sites. "As the manager of the affiliate network, LeadClick had a responsibility to ensure that the advertisements produced by its affiliate network were not deceptive or misleading." I thought the legal standard required
      "ability," but the court tautologically uses the term "responsibility" to satisfy this element. Also note that the court's legal standard ("has the authority to control the deceptive content at issue") sounds a lot like principal-agency liability, but the court doesn't say or imply that LeadClick had a principal-agency relationship with affiliates. Apparently the court is applying some kind of agency-lite liability.

    Finally, the court says that LeadClick's intent to deceive consumers is irrelevant; "it is enough that it orchestrated a scheme that was likely to mislead reasonable consumers."

    Section 230

    Because of the court's intellectual corner-cutting that LeadClick committed a "direct" violation of the FTCA, the Section 230 immunity was already doomed. This is consistent with the Neovi case, where Section 230 didn't even come up even though all of the fraudulent content was provided by third parties. Even though Section 230 doesn't apply to a defendant's own legal violations, the court unfortunately decides to muck up Section 230 jurisprudence anyway, apparently for kicks.

    I believe this is only the second time that the Second Circuit has discussed Section 230. The prior case was GoDaddy's undramatic 2015 win in Ricci v. Teamsters, issued per curiam. Oddly, this panel doesn't cite the Ricci case at all -- not even once. The opinion simply says "We have had limited opportunity to interpret Section 230" without referencing the Ricci case by name. I'm baffled why this opinion so deliberately avoided engaging the recent and obviously relevant Ricci precedent…? Could it be that Ricci would have forced the panel to reach a different result or clearly created an intra-circuit split? Is there some kind of behind-the-scenes politics among Second Circuit judges? I welcome your theories.

    The court runs through the standard 3 prong test for Section 230's immunity:

    1. provider/user of an interactive computer service (ICS). The court correctly says "Courts typically have held that internet service providers, website exchange systems, online message boards, and search engines fall within this definition." (What is a "website exchange system"?). Then the court goes sideways, saying it is "doubtful" that LeadClick qualifies as an ICS because it acts as an affiliate manager that doesn't provide access to servers.

      LeadClick argued that it provided affiliate tracking URLs and recorded activity on its server, but the panel responds that LeadClick didn't cite any cases applying Section 230 in similar contexts. The court continues that LeadClick's tracking service "is not the type of service that Congress intended to protect in granting immunity" because "routing customers through the HitPath server before reaching LeanSpaʹs website[] was invisible to consumers and did not benefit them in any way. Its purpose was not to encourage discourse but to keep track of the business referred from its affiliate network."

      Say what? Affiliate programs are just another form of advertising, so like other advertising programs, they help compensate publishers for creating and disseminating their content. We may not want this particular content (fake news sites touting dubious weight loss products). Even so, affiliate programs do support discourse, and the court's denigration of affiliate programs' speech benefits is unfortunate and unsupportable. More generally, the court seems to be marginalizing the speech benefits that third party vendors give to publishers, which is obviously misguided when vendors help publishers conduct their business more efficiently. I hope other courts don't apply a "discourse promotion" threshold for applying Section 230.

      We rarely see cases turn on the ICS prong, so it's really shocking to see the court go there -- especially when it eventually expressly punts on the issue, making this discussion dicta.

    2. content provided by another information content provider (ICP). The court cites Accusearch for the proposition that ICP "cover[s] even those who are responsible for the development of content only in part,ʺ but then adds a "defendant, however, will not be held responsible unless it assisted in the development of what made the content unlawful."

      The court says LeadClick "participated in the development of the deceptive content posted on fake news pages" because it recruited affiliates knowing some had fake news sites, paid them, occasionally advised them to edit content, and bought ads on legitimate news sites. In other words, the court cites the exact same evidence of LeadClick's prima facie liability as evidence of its lack of qualification for Section 230. This is just another way of saying that once the Second Circuit treated LeadClick as a direct violator of the FTCA, LeadClick had no chance of qualifying for Section 230.

      Notice that none of the cited facts actually involve content "creation" by LeadClick, so the court apparently assumes content "development" covers other activities -- but doesn't say what that term means.

      The court continues: "LeadClickʹs role in managing the affiliate network far exceeded that of neutral assistance. Instead, it participated in the development of its affiliatesʹ deceptive websites, ‘materially contributing to [the contentʹs] alleged unlawfulness.'" What does "neutral assistance" mean, and how does that relate to Section 230 immunity? I assume all future plaintiffs in the Second Circuit will claim that the defendant provided "assistance" to the content originator that wasn't "neutral." That should be fun.

    3. treated as publisher/speaker. The court pulls the same trick with this prong, i.e., LeadClick was facing direct liability due to its own misconduct and citing evidence from the prima facie case as disqualifying evidence for this prong.

    Further Implications

    As we all know, no business wants to litigate against the FTC in court. Not only do the FTC's litigation resources dwarf those available even to large defendants, but judges give the FTC extra credit as the voice of consumers. This case highlighted how the Second Circuit bent plenty of legal doctrine to get the FTC its win. Future defendants who want to fight the FTC in federal court, take note. This kind of doctrinal distortion happens far too frequently in FTC cases, so it would be a mistake to treat it as an unlikely-to-repeat accident.

    There is so much unnecessary bad stuff here for Section 230 jurisprudence in the Second Circuit. Plaintiffs can find plenty of mischief in the court's discussion about what qualifies as "interactive computer services," "neutral assistance" and "development." Yuck.

    In a footnote, the court says the analysis would be the same under Connecticut's UTPA. This suggests that state AGs could similarly establish a prima facie "direct" violation against defendants like LeadClick per their state unfair competition laws without running afoul of Section 230 either. I expect we'll see this case cited extensively by state AGs in future enforcement actions.

    Section 230's year-of-woe keeps going. I'm ready for 2016 to be over. Perhaps the Section 230 pendulum will swing back towards defendants in 2017.

    Republished from Eric Goldman's Technology & Marketing Law Blog

    Read More | 10 Comments | Leave a Comment..

    Posted on Techdirt - 12 August 2016 @ 1:06pm

    FTC Sues 1-800 Contacts For Restricting Competitors From Using Competitive Keyword Advertising

    from the contact-lens-fight dept

    This is a crosspost from Professor Eric Goldman's website.

    For over a decade, I've blogged about 1-800 Contacts' campaign to suppress competitive keyword advertising, including its legislative games (e.g., those times when 1-800 Contacts asked the Utah legislature to ban competitive keyword advertising) and at least 15 lawsuits against competitors costing millions of dollars in legal fees. I've also marveled at its duplicity; 1-800 Contacts historically employed the same competitive keyword advertising practices it subsequently sought to suppress.

    Things have been quiet on the 1-800 Contacts front for the past several years, after it suffered a major blow in the 10th Circuit's ruling, but sometimes the machinery of justice keeps turning quietly in the background. This week, the FTC sued 1-800 Contacts for antitrust violations. I believe this is the FTC's first foray into keyword advertising issues, and it's left some folks scratching their heads.

    The FTC's Allegations

    Let's take a closer look at the FTC's allegations. (As you know, pleadings by government entities are usually a mixture of truth, half-truth and fiction). The complaint says 1-800 Contacts has 50% share of the online retail market for contact lenses. Facing emerging competition from lower-priced entrants, starting in 2004, 1-800 Contacts pursued trademark enforcement against advertisers engaged in competitive keyword advertising. 14 advertisers agreed to settlement terms; only didn't give in. (The complaint redacts the names of the settling advertisers, but in a supplement to this blog post, I identify many of them). The settlement agreements barred the competitors from bidding on "1-800 Contacts" or variants; and 1-800 Contacts reciprocally agreed not to bid on the competitors' trademarks.

    13 agreements also required the competitors to put "1-800 Contacts" on the negative keyword list. Thus, a search for "1-800 Contacts Cheaper Competitors" -- which strongly implies that the consumer sought competitive alternatives to 1-800 Contacts -- would not display these competitors' ads. The FTC provides a screenshot of that search, shown at the right. As you can see, 1-800 Contacts is the only keyword advertiser. However, the top organic result is a link to, which claims to be "an unbiased source of trustworthy information on eye health and vision correction options" and possibly would be a helpful aggregator for consumers. The organic search results also link to some deep pages on 1-800 Contacts' site, none of which consumers would find responsive to the query. While these search results seem a little funky, the search term "1-800 Contacts Cheaper Competitors" is a VERY long-tail query that few, if any, consumers actually tried. The settlement agreements had reciprocal negative keyword obligations for 1-800 Contacts.

    The complaint alleges that "1-800 Contacts has aggressively policed the Bidding Agreements, complaining to competitors when the company has suspected a violation, threatening further litigation, and demanding compliance."

    1-800 Contacts' campaign to restrict competitive keyword advertising could potentially hurt three different marketplace players: (1) the competitors who are hamstrung in their efforts to reach interested consumers, (2) consumers who suffer from a less competitive market, and (3) search engines whose ad auctions are rendered less efficient (and less profitable) when interested bidders choose not to participate. The complaint recaps some of the harms allegedly caused by 1-800 Contacts' conduct, including:

    • distorting the price-setting mechanisms of search engine ad auctions
    • degrading the quality of search results pages by keeping them from displaying the most relevant ads to consumers
    • preventing truthful non-misleading information from reaching consumers
    • suppressing price and service competition among online contact lenses retailers, which causes "at least some consumers to pay higher prices for contact lenses than they would pay absent the agreements, acts, and practices of 1-800 Contacts"
    • increasing consumer search costs to purchase contact lenses online

    In other words, the FTC sees competitive keyword advertising as contributing to efficient search advertising auctions and, more importantly, improving consumers' choices and fostering vendor competition on price and quality. Viewing competitive keyword advertising as pro-competitive isn't novel, but it's satisfying to see the FTC embrace the view so enthusiastically.

    Finally, the complaint enumerates the FTC's wish list of remedies, including:

    • banning 1-800 Contacts from entering into contracts with competitors that restrict participation in search engine ad auctions or suppress the dissemination of truthful non-misleading information
    • banning 1-800 Contacts from "filing or threatening to file a lawsuit against any contact lens retailer alleging trademark infringement, deceptive advertising, or unfair competition that is based on the use of 1-800 Contacts' trademarks in a search advertising auction. Provided, however, that Respondent shall not be barred from filing or threatening to file a lawsuit challenging any advertising copy where Respondent has a good faith belief that such advertising copy gives rise to a claim of trademark infringement, deceptive advertising, or unfair competition."

    While the FTC's focus on keyword advertising is new, its interest in advertising restrictions is not. For example, our Advertising and Marketing Law casebook covers the FTC v. Polygram case from 2003, in which the FTC successfully pursued two competitors' agreements to restrict advertising of old stock in order to prop up a new product release.

    Questions Raised

    Why Is the FTC Acting Now? The FTC says 1-800 Contacts started its enforcement-and-settlement campaign in 2004. Why is the FTC acting now, a dozen years later?

    Normally a complaint like this is instigated by a competitor's complaint, and it would make sense if tipped off the FTC about its situation. (In addition to the trademark battle, had a parallel antitrust lawsuit against 1-800 Contacts going back years). However, I assume would have raised this issue with the FTC a long time ago. After all, filed its antitrust lawsuit in 2011. Perhaps the FTC waited to see how that lawsuit would play out before deciding whether or not to intervene. The district court dismissed's antitrust complaint in 2014; I then see a notice of appeal to the 10th Circuit but it's murky what happened after that.

    Perhaps the FTC is acting now because competitive keyword advertising law has cleared up a lot over the years. As I've mentioned many times, lawsuits over buying competitors' trademarks haven't succeeded in court for about a half-decade; and even lawsuits over the inclusion of a competitor's trademark in the ad copy rarely make much progress in court any more. While I doubt the FTC could have confidently taken a strong stand on the legality of competitive keyword advertising in the aftermath of the Second Circuit's 2009 Rescuecom opinion, the jurisprudential dust has settled a lot since then.

    It's also possible that the FTC finally appreciated how restrictions on competitive keyword advertising distort ad auctions. Auctions work really well to set market prices when all of the relevant bidders participate. I could see how the FTC Bureau of Competition folks, steeped in economics doctrine, had their interest piqued when they first learned about agreements among potential auctions bidders not to participate in the ad auction. Perhaps it took a while for the issue to find its way to the right folks.

    Will 1-800 Contacts Accept a Big Fight Against the FTC? When the FTC conducts an investigation into a complaint like this and thinks there's a problem, inevitably it discusses settlement options with the investigated company before suing. Therefore, it seems very likely that 1-800 Contacts already refused a settlement offer from the FTC. I can understand why 1-800 Contacts might do so. Presumably, 1-800 Contacts believes that its actions over the past dozen years are justified, and it's willing to throw more money to defend that proposition (and, as discussed below, to try to maintain its above-market prices).

    Still, fighting the FTC is a daunting challenge for any company. The FTC always says it's a small agency, but it's still the freaking U.S. government and has more resources than any company it targets. Further, it has an exceptionally strong batting average in litigation, and judges view the FTC as the voice of the consumer -- making it a more sympathetic litigant than a competitor trying to defend its profitable investments in competitive keyword advertising. Furthermore, the FTC has picked a friendly litigation venue. The FTC steered this case into its in-house adjudication process, so the case will be heard before an FTC administrative law judge with appeals going to the FTC Commissioners before this case can be heard in federal court. By keeping this litigation within the FTC, it will take years and lots of money before 1-800 Contacts can tell its story to an adjudicator not employed by the FTC.

    I respect companies that have the fortitude and wealth to stand up to the FTC, but I often question their wisdom and logic.

    Is Competitive Keyword Advertising Legitimate? The FTC's complaint assumes, but doesn't prove, that competitive keyword advertising is a legally legitimate practice. For example, the FTC alleges (para. 18) that "1-800 Contacts claimed—inaccurately—that the mere fact that a rival's advertisement appeared on the results page in response to a query containing a 1-800 Contacts trademark constituted infringement." I'm sure 1-800 Contacts (and all trademark owners) would love to see the FTC's citations for the "inaccurately" comment. Later (para. 32), the FTC says—again, without any citations—that agreements not to engage in competitive keyword advertising "exceed the scope of any property right that 1-800 Contacts may have in its trademarks, and they are not reasonably necessary to achieve any procompetitive benefit."

    Now, as you know, I emphatically agree with this proposition. I've argued for over a decade that competitive keyword advertising is pro-competitive and should be legal; and I've chronicled the systematic failure of trademark owners' anti-keyword advertising lawsuits over the past half-decade. However, I acknowledge that this issue is still being hotly contested in the courts. Indeed, just last week I blogged about a ruling sending a competitive keyword advertising lawsuit (with the trademark used in the ad copy) to a jury because the defendant couldn't convince the judge that it was entitled to summary judgment. So while I wish the state of competitive keyword advertising law was definitively resolved, the FTC's implied factual claim is aggressive.

    For those of us a little tired of the decade-long competitive keyword advertising battles, the FTC's move offers some tantalizing prospects. Because the FTC stacked the litigation deck in its favor, we could get some clean and powerful judicial pronouncements about the legitimacy and pro-competitive nature of competitive keyword advertising. Combined with developments like the Texas ethics opinion greenlighting competitive keyword advertising by lawyers, this case could help push the pendulum so decisively in favor of competitive keyword advertising that it permanently ends the debates.

    What About Vertical Restrictions on Competitive Keyword Advertising? This case deals with horizontal restrictions between competitors. While those are relatively rare, it's quite common (at least in certain industries) for trademark owners to restrict keyword ad bidding by vertical channel partners such as affiliates and distributors. What implications does this lawsuit have for those vertical restrictions?

    Usually, distributors can use manufacturers' trademarks for the goods or services they resell without a trademark license. In contrast, affiliates usually need a trademark license, in which case the trademark owner should be able to put conditions on its trademark license to affiliates. However, trying to impose those same conditions on distributors could be a legal overreach because they didn't need trademark permission at all.

    The FTC might be signaling that it's a problem to restrict keyword ads by channel partners who don't need trademark permission in the first place. However, manufacturers have substantial power to control intra-channel conflicts (see, e.g., the modern deference to resale price maintenance), and restrictions on keyword advertising help the trademark owner manage the trademark and prevent channel partners from driving up the owner's cost of doing so. So vertical restrictions on keyword advertising bidding may have better competitive justifications than horizontal restrictions. My guess is that the FTC didn't intend to implicate vertical restrictions; but it probably wouldn't categorically greenlight them either because some vertical restrictions could indeed have anti-competitive effects.

    What Does This Mean For Trademark Owners? Trademark owners, PAY ATTENTION. Effectively, the FTC is saying that 1-800 Contacts committed antitrust violations by making overreaching trademark demands. I can't recall when the FTC last implied that trademark overclaiming could create antitrust problems (nothing comes to mind immediately). Among academics, we've frequently discussed how trademark overclaims can hurt competition, so many academics probably think it's about time the FTC moved in that direction. However, if this lawsuit signals that the FTC plans to pay more attention to trademark owner overreaching, that would be a seismic event for the trademark owner community.

    Two related notes. First, as I've said before, I think 1-800 Contacts is an exceptionally weak trademark because it's more a phone number than a source identifier. Just like trademark law won't protect [noun].[tld] domain names (when the domain name relates to the noun), "800 [noun]" also should be generic. It creates a lot of friction when we weaponize such highly descriptive terms, and it makes sense for the FTC to pay particular attention to those weapons deployments.

    Second, I've observed before that owners of weak descriptive marks tend to be the most litigious and make the most aggressive interpretations of trademark law. They often use litigation to try to overcome the intrinsic shortcomings of the mark; and they are often paranoid about the so-called "policing duty" that makes trademark owners think they will lose the mark if they don't shut down other users of the term. However, the trademark policing "obligation" is often overstated; and the TTAB has expressly said there's no policing obligation against competitive keyword advertising.

    Read More | 1 Comments | Leave a Comment..

    Posted on Techdirt - 20 May 2016 @ 2:10pm

    Appeals Court Muddies Trademark Nominative Fair Use Doctrine

    from the clear-as-mud dept

    The nominative use doctrine allows third party references to trademark owners using the trademarks they chose as their preferred descriptors. Without a robust and well-functioning nominative use doctrine, trademark owners can have too much control over their brands -- they can shut down the advertisement of complementary or competitive offerings and potentially even critical scrutiny of the brands. Unfortunately, Congress never adopted a statutory nominative use doctrine for trademark infringement, and the doctrine seemingly baffles the courts. As a result, the circuits have created a patchwork of nominative use doctrines. A ruling this week from the Second Circuit exacerbated this problem.

    The lawsuit itself is so ridiculous that it's hard to discuss with a straight face. The plaintiff owns the certification mark "CISSP" for certifying information security professionals. I've blogged about them before (1, 2). The defendant, Security University (SU), earned the CISSP designation and then ran ads self-referring to itself as "Master CISSP" or a "CISSP Master." When the certification mark owner challenged the "master" reference, the defendant responded "SU will continue to use the word Master. Master Clement Dupuis is a Male Teacher [and] thus he is a Master according to the dictionary." Oh come on, this is off the BS charts. Nevertheless, the district court dismissed the lawsuit because the defendant earned the CISSP certification and therefore could make a nominative use of it.

    Trademark law isn't well-equipped to deal with an issue like this. The defendant's "master" usage implies the defendant has a superior knowledge compared to other CISSP. That's more of a false advertising issue than a trademark issue, and courts struggle when pressing trademark law to regulate broader advertising law issues. Furthermore, the way certified parties advertise themselves is best addressed by the terms of the certification, i.e., the certification should include restrictions on making superlative claims, forming combination marks or otherwise creating the impression that there are new certification flavors. I think contract principles, incorporated through the certification part of a certification mark, would be more efficacious than primary trademark doctrine.

    Still, this case presented itself as a trademark case to the courts, so they address it using the tools of trademark. The Second Circuit reverses the district court's nominative use determination, remanding the case back to the district court to apply the legal principles it lays out in the opinion.

    Regarding the nominative use doctrine, the court makes two major moves. First, the court could have adopted the legal articulation used by other circuits, such as this venerable language from the Ninth Circuit's New Kids on the Block case from a quarter-century ago:

    First, the product or service in question must be one not readily identifiable without use of the trademark; second, only so much of the mark or marks may be used as is reasonably necessary to identify the product or service; and third, the user must do nothing that would, in conjunction with the mark, suggest sponsorship or endorsement by the trademark holder.

    Instead, the court chose to re-articulate the doctrine in its own words:

    (1) whether the use of the plaintiff's mark is necessary to describe both the plaintiff's product or service and the defendant's product or service, that is, whether the product or service is not readily identifiable without use of the mark; (2) whether the defendant uses only so much of the plaintiff's mark as is necessary to identify the product or service; and (3) whether the defendant did anything that would, in conjunction with the mark, suggest sponsorship or endorsement by the plaintiff holder, that is, whether the defendant's conduct or language reflects the true or accurate relationship between plaintiff's and defendant's products or services.

    You can see similarities to the Ninth Circuit test, but notice the differences. The Second Circuit's first factor requires a "necessity" of the trademark reference, a seemingly higher bar than the Ninth Circuit's "not readily identifiable" requirement (a phrase which nevertheless confusingly modifies the word "necessary"). In the second factor, the Second Circuit omits the "reasonably" qualifier from the identification necessity. In the third factor, the Second Circuit requires that the description be "accurate."

    All told, the Second Circuit's modifications make it just a little harder for defendants to qualify for the nominative use case. Why did the Second Circuit toughen up the nominative use standard compared to the Ninth Circuit? The court provides no explanation, but I suspect it's an overreaction to the defendant playing semantic games ("bad facts make bad law," etc.), plus the overall misfit of using trademark law to govern what's really a false advertising issue.

    Second, the Second Circuit addresses when a court should do its nominative use analysis. In the Ninth Circuit, the three-factor nominative use test substitutes for the standard multi-factor likelihood of consumer confusion test. In the Third Circuit, nominative use is an affirmative defense. Rejecting both approaches, the Second Circuit says that, "in nominative use cases," judges should add the three nominative use factors to the standard eight-factor Polaroid consumer confusion test -- in other words, an overlong eight-factor test now becomes an even longer 11-factor test. Any ideas how that's going to work out? (Recall Judge Walker's cogent and pithy critique: "Every circuit uses some n-factor test for likelihood of confusion, where n is an integer that is almost certainly too large for the task at hand").

    By adding factors to the multi-factor test, the court makes it effectively impossible for defendants to establish nominative use on a motion to dismiss (a possibility under the Ninth Circuit's approach, even if unlikely), and it will inhibit judges from ruling for the defense on summary judgment. Judges can resolve consumer confusion analyses on summary judgment, but some judges are reluctant to do so because the factors encode so many factual questions. Adding three more factors to the test, each with additional factual predicates, will make judges that much more reluctant. So the court's ruling implicitly adds to the costs of a successful nominative use defense.

    As I mentioned before, courts can't figure out how to deal with the nominative use doctrine. This is a problem because the nominative use doctrine plays a critical role in the successful functioning of markets, something I emphasized in my 2005 Deregulating Relevancy article. To facilitate socially beneficial nominative uses, we need doctrinal and procedural fastlanes that provide certainty and cheap resolutions to secondary users. The Second Circuit decision moves us further away from that direction.

    A nomenclature note: there is some disagreement about whether to call the doctrine "nominative use" or "nominative fair use." The Second Circuit opinion uses the term interchangeably but seems to prefer "nominative fair use" -- that gets 34 references in the opinion, compared to 11 for "nominative use."

    Reposted from the Technology & Marketing Law Blog

    Read More | 1 Comments | Leave a Comment..

    Posted on Techdirt - 19 July 2013 @ 4:14pm

    Once Again, Buying Competitive Keywords Is Not Trademark Infringement

    from the killing-initial-interest-confusion-doctrine dept

    Here are some things I hate:

    • duplicitous litigants, such as plaintiffs who buy competitive keyword advertising yet sue competitors for doing the same thing
    • economically irrational and socially wasteful litigation, such as plaintiffs who spend over a million dollars in legal fees on a problem that, at most, is worth tens of thousands of dollars
    • pugnacious litigation over incredibly weak trademarks, like so-called trademarks that are simply 1-800 plus a generic noun
    • trademark doctrines that have no rigorous definition or scientific support, like the initial interest confusion doctrine
    • technologically unsavvy judges that guess about how consumers use the Internet when those assumptions have been proven false in the literature
    • appellate courts that eviscerate a legal doctrine without expressly saying the doctrine is dead, leaving the legal doctrine as a zombie to plague the courts for many years
    (Why yes, I did wake up on the wrong side of the bed this morning).

    With this list in mind, you can see why I hate the 1-800 Contacts v. lawsuit. 1-800 Contact has spent enormous amounts on legal fees—at least $650k as of 2010--pursuing for competitive keyword ads that had generated $20 in profit for (no, that's not a typo) and, at maximum, a few tens of thousands of dollars in revenue for affiliates. All of this litigation is predicated on the initial interest confusion doctrine, an overly amorphous doctrine that no one can define or find any scientific support for, and which has been has been a loser in court for many years. To top it off, 1-800 Contacts had hypocritically engaged in competitive keyword advertising itself. Glass house and stones, etc.

    After 6 years in court, the case isn’t over yet. This week, the Tenth Circuit affirmed most of the district court’s opinion and emphatically rejected most of 1-800 Contacts' lawsuit against for the competitive keyword advertising it and its affiliates did. However, a small issue got remanded for a jury trial, so the parties will get the pleasure of wasting many tens of thousands of dollars more to conduct the jury trial unless they can finally find a way to settle. Hooray for litigation that create massive deadweight social losses!

    Background bought AdWords keyword advertising on the following keywords: “1-800 contact lenses”; “1800 contact lenses”; “800 contact lenses”; “”; “”; “”; “”; “”; and “” In addition, had 10,000 affiliates through Commission Junction. Two affiliates, Dusty Goggans and Ryan McCoy, bought the keyword “1800Contacts” and similar terms; and McCoy ran at least one ad that displayed “1800 Contacts” in the ad copy.

    1-800 Contacts sued in 2007. In 2010, the district court destroyed 1-800 Contacts' arguments, handing a complete win. On appeal, the Tenth Circuit upholds most of the lower court's ruling.

    Is Buying Competitive Keyword Advertising Trademark Infringement?

    1-800 Contacts argued that created initial interest confusion by buying 1-800 Contacts' trademarks. The Tenth Circuit adopted one of the nation's most favorable definitions of initial interest confusion in its horrible 2006 case Australian Gold v. Hatfield. That case said initial interest confusion “results when a consumer seeks a particular trademark holder’s product and instead is lured to the product of a competitor by the competitor’s use of the same or a similar mark.” If this is just a restatement of bait-and-switch law, fine, but it’s generally believed the Tenth Circuit standard covered more than bait-and-switch. Bait-and-switch requires an initial falsity to bait the hook but the 10th Circuit's standard didn't mention falsity, and there’s no intrinsic falsity in a typical competitive keyword ad buy where the competitor's trademark is used to trigger ad copy.

    By embracing the initial interest confusion in 2006, the Tenth Circuit was a late adopter to that party. Since then, the Ninth Circuit pivoted away from initial interest confusion doctrine in the Network Automation case (cited in this opinion), and other courts have cast doubts on the doctrine.

    Unfortunately, the Tenth Circuit doesn't have the fortitude to knock out the doctrine entirely. Instead, it offers this broad legal interpretation:
    initial-interest confusion would arise as follows: a consumer enters a query for “1-800 Contacts” on Google; sees a screen with an ad for that is generated because of’s purchase of one of the nine Challenged Keywords; becomes confused about whether is the same source as, or is affiliated with, 1-800; and therefore clicks on the ad to view the site. has exploited its use of 1-800’s mark to lure the confused consumer to its website.
    Normally I would strenuously object to this standard, but the Tenth Circuit then did a tricky sleight of hand with its evidentiary standards. Typically, trademark law measures consumer confusion through consumer surveys asking for their confusion about product source. Cases have usually held that 20%+ consumer confusion in those surveys is probative of confusion while <10% consumer confusion is probative of no legally actionable confusion.

    Here, the court says that clickthrough rate (CTR) is a proxy for a consumer confusion survey--and the 10%/20% presumptions apparently apply. What this means in practice is that if the ad doesn't get well over a 10% clickthrough rate, then the low clickthrough rate will be probative of an absence of consumer confusion. But NO ad EVER gets 10%+ clickthrough rate (and certainly nowhere close to the 20% rate that would be decisively plaintiff-favorable); it’s big news when an ad clears 3% CTR.

    With respect to the ads without the 1-800 Contacts trademark in the ad copy, got a 1.5% CTR and its affiliates got less than 1% CTR (0.6-0.7%). Because the CTR was dramatically below the 10% threshold that's probative of a lack of consumer confusion, there was no actionable initial interest confusion either by or its affiliates.

    So this evidentiary standard should be dispositive in other initial interest confusion cases. If other courts follow the same logic, no competitive keyword ad could ever possibly create initial interest confusion because even the best performing ad’s CTR will be below the legally recognized threshold of consumer confusion.

    Naturally, if plaintiffs can find other evidence to support initial interest confusion, I don't think this ruling precludes those arguments. But since no one knows what initial interest confusion is, and defendants can always introduce their single-digit CTR, I think we're done with the argument that keyword ads create initial interest confusion, at least in the Tenth Circuit.

    Unfortunately, because the Tenth Circuit didn't expressly kill the doctrine outright, it's going to keep appearing in plaintiff arguments for many years hence. That means parties will waste a lot more money in litigation because the Tenth Circuit couldn't just say what it meant. SIGH.

    In addition to grousing about the zombie-fication of the initial interest confusion, I have to kvetch about this passage from the opinion (emphasis added):
    even if consumers in general may not much care what retailer supplies their contact lenses, the consumers relevant to this suit are looking for a particular retailer. Presumably they have narrowed their search because they have already selected 1-800 as the preferred retailer and are searching for its website or perhaps commentary on its performance. Given the purpose of the search, the shoppers will be attentive to click on those results that will connect them with sites relating to 1-800. In addition, once the consumers see the results page, the substantial dissimilarity between “1-800 Contacts” and “” (or its other websites) can be expected to greatly reduce the chance that the consumers will think that the parties are related enterprises; the similarity of the search term and 1-800’s mark is of minor relevance.

    Perhaps in the abstract, one who searches for a particular business with a strong mark and sees an entry on the results page will naturally infer that the entry is for that business. But that inference is an unnatural one when the entry is clearly labeled as an advertisement and clearly identifies the source, which has a name quite different from the business being searched for.
    Kudos to the Tenth Circuit for recognizing that ad copy educates consumers and helps searchers improve their search process. Big demerits, however, for assuming that a search for "1-800 Contacts" signals that the searchers are, in fact, looking for 1-800 Contacts (or even information about it). This is demonstrably false, as the Hyman/Franklyn study showed and as I explained in 2005. The Network Automation opinion made this identical error. Why are appellate judges so tempted to assume searcher intent from the search keyword?

    Contributory Infringement for Affiliates' Ad Buys With the Trademark in the Ad Copy

    Although the court's disposition of the initial interest confusion resolves the ads triggered by 1-800 Contact's trademarks that don't display the trademarks in the ad copy, there still remains the ads by's affiliate which contained the trademark in the ad copy. In a discussion that goes on way too long, the court ultimately rejects's vicarious trademark liability, an agency-based doctrine. This should have been an easy call because there's absolutely no way affiliates mediated through Commission Junction are's agents. The court notes the possibility that Commission Junction was's agent and the affiliates were sub-agents. Give me a break. Still, the court says that the affiliate displaying the trademark in the ad copy acted outside the scope of any agency. The court says:
    The issue is not whether McCoy had authority to act on’s behalf at all, but merely whether he had actual authority to publish an ad displaying a variation of 1-800’s mark in its text....McCoy never believed, reasonably or otherwise, that authorized him to place the ads. Thus, the subjective component of actual authority was absent.
    Nice save of a completely faulty premise.

    Nevertheless, the appellate court reopens the contributory infringement issue. The court is bothered by the fact that waited 3 months between when it got 1-800 Contacts' lawsuit and asked the Commission Junction affiliate (McCoy) to stop advertising with 1-800 Contacts' trademark in the ad copy. The court says:
    if could have stopped the use of ads using 1-800’s mark by simply requiring CJ to send an email blast to its affiliates forbidding such use, then’s failure to proceed in that manner after learning of such ads could constitute contributory infringement.….When modern technology enables one to communicate easily and effectively with an infringer without knowing the infringer’s specific identity, there is no reason for a rigid line requiring knowledge of that identity, so long as the remedy does not interfere with lawful conduct.
    This standard doesn't ensure a 1-800 Contacts jury win. It's not clear to me that could have fixed the problem just by sending an email. Still, even if loses this portion of the case, this is a small fraction of the overall conduct 1-800 Contacts was complaining about. I believe the damages exposure for the remaining ads in question is relatively small. It would be more rational for the parties to settle up rather than spend the money to take this question to the jury, but nothing about this litigation has been rational to date, so I'd be surprised if rationality suddenly made an appearance now.


    I must sound like a Johnny One Note, but I'll say it again. Keyword advertising litigation is dead, dead, DEAD. It may not have been dead when 1-800 Contacts brought the lawsuit in 2007, but the intervening 6 years have proven that keyword advertising lawsuits are not easy to win and are economically irrational.

    In this case, it's been clear all along that 1-800 Contacts' wasn't suing because it wanted to protect its brand. Instead, its apparent motivation all along has been to shut down competitors, by draining them of cash if necessary. I'm sure both parties have spent well over $1M on this litigation--money that the incumbent (1-800 Contacts) can afford more easily than newer entrants (like Thus, it's devastating to that both courts have denied its attorneys' fees. Given that 1-800 Contacts got a small victory on appeal, the fee denial is consistent with trademark law, but the consequence is that 1-800 Contacts can get away with lawfare for anti-competitive objectives.

    And 1-800 Contact’s fear of competition has manifested itself in other ways, including similar suits against other competitors and repeated attempts to game its state legislature. did bring an antitrust claim against 1-800 Contacts for this litigation; that case has been on hold awaiting this Tenth Circuit ruling. We'll have to see if now can make any headway to hold 1-800 Contacts accountable for its legal choices.

    Prior Blog Posts on This Case I’ve also blogged numerous times on 1-800 Contacts.

    Related Blog Posts

    For more on keyword advertising legal developments, see my recent Forbes posts on trademark infringement and keyword advertising: Also see my two related law review articles:

    Read More | 9 Comments | Leave a Comment..

    Posted on Techdirt - 12 October 2012 @ 12:29pm

    The 'Cloud Computing Act Of 2012'... Or How Internet Regulation Can Go Awry

    from the paved-with-good-intentions dept

    Sen. Amy Klobuchar has introduced a new bill, the "Cloud Computing Act of 2012" (S.3569), that purports to "improve the enforcement of criminal and civil law with respect to cloud computing." Given its introduction so close to the election, it's doubtful this bill will go anywhere. Still, it provides an excellent case study of how even well-meaning legislators can botch Internet regulation.

    What the Bill Does

    From its 1980s origins as a law restricting hacking into government computers, the Computer Fraud and Abuse Act (CFAA) has morphed into a general-purpose federal law against trespassing on anyone else's computers. With that breadth, the CFAA extends to a wide variety of activities, ranging from data scraping (see, e.g., EF Cultural Travel v. Explorica) to fake profiles (see, e.g., the Lori Drew prosecution related to Megan Meier's death) to ex-employees walking out the door with competitively sensitive information (see, e.g., US v. Nosal and WEC v. Miller).

    The proposed bill's main substantive provisions attempt to give "cloud computing services" extra protections under the CFAA. First, the bill says that each unauthorized access of a cloud computing account counts as a separate CFAA offense. Second, the bill specifies a formula for computing losses in CFAA violations involving cloud computing services, setting a minimum floor of $500 loss per affected cloud computing account.

    Problems with the Bill 

    The CFAA is Already a Mess.  Good luck trying to read the CFAA's text. Constant amendments over the years have created spaghetti code. This bill adds only slightly to the CFAA's overall lack-of-tidiness, but every incremental amendment makes the CFAA more unwieldy.

    The Definition of "Cloud Computing Service" is Incoherent. The bill seeks to protect cloud computing services, but what are those? Check out the bill's definition:

    the term "cloud computing service" means a service that enables convenient, on-demand network access to a shared pool of configurable computing resources (including networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or interaction by the provider of the service.

    What??? This sounds more like a vendor's sales pitch than a basis for criminal prosecution. We can reinforce the definition's weakness by trying to determine what isn't a cloud computing service. Every user-generated content website seems to qualify; but so should every online bank. In fact, this definition of cloud computing service probably becomes co-extensive with the Internet generally.

    To be fair, the failed definition isn't totally the drafter's fault. I don't think it's possible to define "cloud computing service" precisely. Tip to legislators: if you can't clearly define your subject matter of your legislation, you're probably doing something wrong.

    What's the Problem That Needs to Be Solved? I can't figure out how the proposed amendments address any problem we're seeing in the field. It's possible I've missed some relevant case, but I can't think of a single case I've seen where the CFAA underprotected a cloud computing service or this legislation would have changed the outcome. Seeking some clarity, I submitted a press inquiry to Sen. Klobuchar's office last week and got no response. So I have no idea what problem this bill purports to solve.


    This bill exemplifies several ongoing problems with efforts to legislate the Internet:

    1) Legislative grandstanding. It's flashy for legislators to tell their constituents that they are fighting hard to protect emerging technologies like "cloud computing." But legislators rarely understand cutting-edge technologies, and usually rapidly evolving technologies are poor candidates for legislative intervention. So legislators' efforts to push buzzword-laden legislation are often more for show than substance.

    2) Regulatory exceptionalism. As I explain here, legislators keep creating new "exceptionalist" rules for subsets of the Internet ecosystem--online dating sites, social networks, cloud computing services, etc. We saw how well that worked in California's effort to ban employers from asking employees for social media login credentials. California so utterly failed at defining "social media" that it simply covered the entire Internet...and all non-networked electronic data too! Yet, legislators seemingly haven't learned from their colleagues' repeated failed efforts to precisely define the contours of some Internet subcommunity. The proposed CFAA amendment, and its gibberish definition of "cloud computing service," exemplifies this.

    3) Code proliferation. For every problem, real or perceived, legislators think they can fix the problem with more regulatory code. But the manufacturing of new legal code exacts a toll of its own. This bill increases the CFAA's complexity with minimal or zero commensurate benefit. If Sen. Klobuchar or anyone else really wants to "fix" the CFAA, a good start would be to reduce the law's length, organize it better, and reduce its implications for users' ordinary Internet activity.

    27 Comments | Leave a Comment..

    Posted on Techdirt - 14 August 2012 @ 3:43am

    Citations & Sarcasm: How Gizmodo Got A Defamation Lawsuit SLAPPed Down

    from the an-opinion-is-not-defamation dept published an article, Smoke & Mirrors: The Greatest Scam in Tech, about Redmond's venture, Peep Telephony. In addition to using the word "scam" in the title, the article had lots of denigrating things to say about Peep and about Redmond's prior initiatives. (The opinion lays out the beefs, although some of the hot spots are apparent from a quick review of the initial article). Gizmodo subsequently published Redmond's rebuttals. Later, Redmond apparently decided the rebuttal wasn't enough and asked Gizmodo to remove both articles, which Gizmodo declined to do. Redmond then sued Gizmodo's parent Gawker Media for defamation. The court dismissed the case on anti-SLAPP grounds, and that means Redmond will owe a check to Gawker for his lawsuit.

    The court has no problem finding that Peep Telephony's activities were a matter of public interest, as Peep Telephony had received some high-profile coverage from technology reporters before Gizmodo's story, and Redmond apparently had been trying to stir up press coverage in advance of the 2011 CES conference. The court summarizes that the "Gizmodo article was a warning to a segment of the public--consumers and investors in the tech community--that Redmond's claims about his latest technology were not credible."

    The court also says that Redmond's beefs relate to statements of opinion, not fact. The court notes that the word "scam" as not a factual assertion (a dicey outcome), the article was written in a "casual" and "sarcastic" first-person style ("the article's general tenor and language would give a reasonable reader the impression the authors were expressing subjective opinions, not reporting facts"), and the article used weasel words, such as "seems," "arguably," "looks like," etc., to qualify key fact-like assertions.

    The most interesting part of the opinion is where the court talks about the article's "transparency." The court says (emphasis added):

    The sources upon which the authors rely for their conclusions are specified, and the article incorporates active links to many of the original sources--mainly Web sites and promotional material created and maintained by Redmond and his ventures....Having ready access to the same facts as the authors, readers were put in a position to draw their own conclusions about Redmond and his ventures and technologies....Statements are generally considered to be nonactionable opinion when the facts supporting the opinion are disclosed.

    This is true, of course, but a point often lost when defamation plaintiffs are breathing fire. A properly-cited article, filled with hyperlinks to original source materials, should be extra-resistant to defamation claims--even if written with typical blogger snark. Readers can easily inspect the source materials themselves and make their own judgments about the article's veracity. Thus, either the citations provide proper factual support for the article's opinion, or the links should eliminate any problems with the author's knowledge (where that matters to the prima facie defamation claim, which would have been the situation here). Either way, the defamation claim should fail, as it did here.

    So this decision is a great ruling for bloggers. Unfortunately, it's unpublished (like far too many California appellate court opinions), which limits its precedential effect. To fix this, my RA and I are planning to request that the court publish it. Even if it remains unpublished, perhaps the ultimate takeaway--that defamation claims against well-cited blog posts will be quickly dismissed by anti-SLAPP laws and lead to the plaintiff paying money to the defense--will help dissuade similar lawsuits nonetheless. Especially in a situation like this, where the potential plaintiff already had gotten an on-the-spot rebuttal, suing over a blog post like Gizmodo's rarely makes sense.

    Read More | 16 Comments | Leave a Comment..

    Posted on Techdirt - 30 March 2012 @ 4:05pm

    No, Telling Customers Why Your Product Is Better Than A Competitor's Is Not Trademark Infringement

    from the and-judges-should-know-that dept

    Every time I read an opinion like this ruling in AR Pillow Inc. v. Cottrell, a little piece of me dies. This is a ridiculously easy case, yet somehow it got all tangled up.

    The litigants are rival vendors of similar products to combat baby acid reflux. For a slightly similar dispute involving trademark fights over baby products, see BabyAge v. Leachco.

    The plaintiff is largely complaining about text on the defendant's website explaining why the plaintiff's product isn't as good as the defendant's. From my perspective, the explanation doesn't constitute a trademark "use" at all because the trademark is being used as a referent. See, e.g., Naked Cowboy v. CBS, 2012 WL 592539 (S.D.N.Y. Feb 23, 2012). Thus, the court should dismiss the trademark claim for non-trademark use. But if the court doesn't do that, at least it should dismiss as a nominative use. See, e.g., 1 800 GET THIN v. Hiltzik. There is absolutely no question that the defendant's reference qualifies as a nominative use under Ninth Circuit law. Yet, and here's where a piece of me dies, the opinion doesn't discuss nominative use AT ALL. What???

    Instead of addressing the two most obvious grounds, the court engages in doctrinal contortions to fit this case into a standard likelihood of consumer confusion (LOCC) analysis. As I've explained elsewhere, the multi-factor LOCC test simply makes no sense when a third party is using the trademark editorially as a referent. My paradigmatic example is the Ballysucks case, where the LOCC analysis is ridiculous because the court is trying to compare a vendor with a griper who registered a "sucks" domain. The LOCC test doesn't work any better here. The judge almost seems to know that the LOCC test isn't the right test, but he doesn't seem to know what else to do.

    Fortunately, the judge overcomes his shaky analysis by reaching the right result, concluding that there's insufficient evidence of consumer confusion. Yay for good outcomes. But we need to find better ways to make it clear trademark law should not play a role in situations like this. For more on that, see my Online Word of Mouth paper.

    One more point: the court's technological discussions are a mixed bag. On the plus side, citing Network Automation and Matt Cutts' blog post/video, the judge rejects the plaintiff's claims over keyword metatags because Google ignores them. But then the judge launches these groaners:

    "there is no evidence in the record that use of the term AR Pillow in Google or other search engine currently leads to defendant's website"

    and later

    "Consumers who search for AR Pillow today are not presented with defendant's website in the rankings"

    Oh god, not this again. Please, let's kill this meme RIGHT NOW. Relative placement of search engine results is a HORRIBLE way to evaluate trademark disputes. First, as I explain here, the junior user doesn't control placements--the search engines do. Second, in my post on the Bitchen Kitchen litigation, I explain technological reasons why this is a terrible idea, including search results personalization, the fact that results change minute-by-minute, and the fact that different search engines rank their results differently. Judges, I beg you, please don't go down this wormhole.

    8 Comments | Leave a Comment..

    Posted on Techdirt - 24 January 2012 @ 7:34pm

    Protip: Make Sure Your Facebook Photos Don't Contradict Your Statements In Court...

    from the just-saying dept

    In my long-running series of litigants saying one thing in court and another when talking to their friends online, consider this from a worker's comp case after a refrigerator fell on Clement:

    The first issue Clement raises in his brief is whether pictures of him that appeared on Facebook and MySpace should have been admitted into evidence. He complains that the pictures are a disgrace to the dignity of the workers' compensation proceedings and the legal system and have nothing to do with his medical treatment....
    We find no abuse of discretion in the allowance of the photographs. Clement contended that he was in excruciating pain, but these pictures show him drinking and partying. Certainly these pictures could have a bearing on Clement's credibility, albeit a negative effect that Clement might not wish to be demonstrated to the ALJ or the Commission. We hold that there was not an abuse of discretion in allowing the photographs.

    Now that Facebook can do facial recognition, it should next develop a tool to detect photos depicting alcoholic drinks and give users a way to automatically opt-out of those photos!

    Prior blog posts in this series:

    * YouTube Video Impeaches Witness' Credibility--Ensign Yacht v. Arrigoni
    * Facebook Entries Negate Car Crash Victims' Physical Injury Claims
    * Contrary MySpace Evidence Strikes a Litigant Again--HAC, Inc. v. Box
    * MySpace Postings Foil Another Litigant--Sedie v. U.S.
    * Disturbingly Humorous MySpace Posts Used as Impeaching Evidence in Spousal Abuse Case--Embry v. State
    * Latest Example of Social Networking Site Evidence Contradicting In-Court Testimony--People v. Franco

    23 Comments | Leave a Comment..

    Posted on Techdirt - 16 January 2012 @ 11:03pm

    I'm Not A Fan Of This Craptastic Trademark Lawsuit

    from the cast-of-fans dept

    We've seen some pathetic trademark lawsuits this year (SUE MOAR KALE, anyone?), but I'll nominate this long-running litigation money-sink (going over 3.5 years) as the saddest trademark case of 2011.

    Fancaster registered its mark in 1989 for broadcasting services, and over the years it's been used in connection with a range of services, "including selling Fancaster branded radios, charging customers to watch closed circuit boxing matches, producing karaoke shows, transmitting sponsored news messages to wireless pagers and cell phones, and conducting live demonstrations of FANCASTER broadcast services" (cites omitted).

    In 2006, it launched to broadcast short sport-related video clips. It hoped to cover such must-see events "as La Tomatina in Spain, Ostrich racing in Arizona, the Westminster Kennel Club Dog Show and the annual Nathan's Hot Dog Eating Contest." Rather than advertise the website on the Internet (you know, where people who enjoy content online might already be), they were seeking out untapped Internet enthusiasts by "marketing the website at sporting events, bars, on local television channels in Sioux Falls, South Dakota and Sioux City, Iowa, on radio stations in Charleston, South Carolina, and via flyers and handbills."

    Meanwhile, in 2008 Comcast rolled out a service called "that allowed users to watch full-length premium mainstream media over the Internet." The service was a debacle, losing $80M in less that 2 years due to “the unexpectedly high cost of distributing video content on the internet.” (Even though Comcast acquired bandwidth at wholesale rather than retail costs... how much would it have cost non-carriers to launch competitive services?). In March 2011, Comcast shut down the Fancast service and rolled the domain name over to XfinityTV.

    With the overlap between the Fancaster and Fancast names, one possibility is that Comcast blatantly ripped off the name of a small startup who wouldn't want to tangle with a giant, thereby creating "reverse confusion" where everyone thinks first-mover Fancaster infringes second-comer Comcast. But another story equally fits these facts: Fancaster is doing a little trademark trolling, seeking to increase Comcast's $80M of losses by grabbing some gravy for itself. (Some gravy indeed: Fancaster's damages expert thought it would take $73M of corrective advertising to fix Comcast's damage to a brand that has no market awareness outside of Sioux City.)

    It's a sad commentary on our milieu when we can't tell which litigant is bullying the other. Maybe *both* parties are equally imbibing the bullying elixir. Fancaster initially unleashed the litigation hounds, but Comcast responded with a hailstorm of countermoves, including an ACPA counterclaim for a slew of "fancast" domain names Fancaster registered after learning about Comcast's upcoming launch. A lot of lawyers appear to have satisfied their billable hour goals using this case. Yay for free-spending, deep-pocketed clients!

    Trademark Infringement

    The court resoundingly thumped Fancaster's core argument about consumer confusion, miraculously finding a way to twist all of the factors to Comcast's favor. The judge may have cut some analytical corners, but that says the judge simply didn't accept Fancaster's narrative.

    The court specifically rejected the possibility of initial interest confusion, citing 3rd Circuit precedent that basically limited IIC to competitors, because the parties didn't directly compete. The court also dismissed Fancaster's efforts to show overlaps in search engine results, saying "the confusion one encounters on an Internet search engine is a twenty-first century version of that experienced when searching the phone book." (I am going to be doing some work this quarter to show that the initial interest confusion doctrine almost never succeeds in court any more, and therefore it imposes costs on both litigants for no gain. This case is just one example of that.)

    The court also scoffed at Fancaster's request for $73M for corrective advertising:

    There is not a shred of evidence of any damage to the Fancaster mark caused by Comcast. The only loss to Fancaster that Mr. Krueger could testify to was that resulting from pursuing the instant litigation against Comcast.

    Evidentiary Issues

    Comcast had survey expert Hal Poret do two surveys. The court tossed the first one because it didn't adequately replicate market conditions by not presenting consumers with a navigable website:

    use of a printout and static screenshots, instead of live websites, provide ample grounds on which to exclude the March 2009 survey. For one, it is difficult to fathom how presenting a respondent with a paper printout of the FANCAST homepage in anyway replicates how an Internet user would encounter and perceive the FANCAST website in the marketplace. Websites, particularly those that offer video content, are meant to be viewed on a computer and allow consumers to browse and interact with them via hyperlinks. The FANCAST printout offered none of these aspects. Similarly, although viewed on a computer, the static screenshots of the fancaster and control website homepages did not allow respondents to interact with them as they ordinarily would in the marketplace.

    I haven't researched this issue, but this ruling may tell us something important about the requirements for consumer surveys when websites are involved.


    This ruling eviscerated Fancaster's case, making it a strong win for Comcast, but it left a few residual legal issues open. Yet, the legal battle has been mooted by the passage of time. Comcast already stopped using Fancast as a brand, and Fancaster still hasn't shown a lot of movement towards developing a real business or even a revenue model. Are the parties really going to spend more money on a pointless lawsuit? We all know what the answer should be; let's see how they actually answer.

    For more on the case, see Rebecca Tushnet's 43(B)log .

    12 Comments | Leave a Comment..

    Posted on Techdirt - 21 December 2011 @ 3:42pm

    Court Indicates Facebook May Be Violating Your Publicity Rights With Sponsored Stories

    from the don't-like-the-service,-but-lots-of-questions-in-the-ruling dept

    Because Facebook does so many things that aren't in users' interests, their "Sponsored Stories" program barely registers. Nevertheless, Sponsored Stories demonstrates why many people are burned out on Facebook. Facebook collects user preferences through its semantically ambiguous "like" button and then uses that data to show ads to the users' friends with a seeming endorsement. Using my preferences does little to advance my relationship with my friends, but the implicit endorsement is designed to get my friends to investigate the ads, increasing the advertiser's credibility and Facebook's profits. So Sponsored Stories creates a zero-sum game: I, as a user, probably don't get any value from the public presentation of my implicit endorsement (if anything, it might hurt my position with my friends), but Facebook and its advertisers benefit from it.

    My response to Facebook's roll out of Sponsored Stories was swift and decisive: I don't "like" any businesses on Facebook or do any other activities on Facebook that I believe can trigger a Sponsored Story. (I would also categorically opt-out of being a part of Sponsored Stories if Facebook actually let me decide what I want to share with my friends, but Facebook doesn't). Instead, if I want to make a commercial recommendation to my friends--something I do occasionally--I just share it directly in my status report. That way, I control the message I deliver to my friends, instead of letting Facebook or advertisers control how they communicate my interest to my friends. And the zero-sum nature of Facebook's offering drives a deeper wedge into my relationship with Facebook, making me less willing to use Facebook generally and more receptive to alternatives.

    To me, this marketplace response is adequate. To plaintiffs' lawyers in Fraley v. Facebook, however, Sponsored Stories gives another reason for a litigation fiesta. Remarkably, unlike so many other "privacy" lawsuits against Internet companies, this lawsuit survives the motion to dismiss--dramatically increasing the odds that Facebook will be writing a check for this so-called "feature."

    This is a rich and interesting opinion by Judge Koh (embedded below) that has something for everyone to "like" (or dislike). Some of the highlights:

    Article III Standing

    In a ruling that bucks a mini-trend, Judge Koh upholds the case from an Article III standing challenge. She says that violation of a statutory right (in this case, California's publicity rights statute) automatically satisfies the actual harm requirement of Article III standing. The plaintiffs also satisfied the "particularized" and "concrete" requirements of Article III by explaining how the Sponsored Stories feature used their information.

    She explicitly distinguishes numerous defense-side Article III wins (including her own recent iPhone application litigation and Low v. LinkedIn decisions) by noting the particular nature of the plaintiffs' publicity rights claim. In this case, unlike the others, the plaintiffs are claiming that their endorsement had commercial value to help sell goods to others, compared to the situation in the prior cases where the commercial value of a user's data came from theoretically improved marketing to the user him/herself. She says:

    Plaintiffs here do not allege that their personal browsing histories have economic value to advertisers wishing to target advertisements at Plaintiffs themselves, nor that their demographic information has economic value for general marketing and analytics purposes. Rather, they allege that their individual, personalized endorsement of products, services, and brands to their friends and acquaintances has concrete, provable value in the economy at large, which can be measured by the additional profit Facebook earns from selling Sponsored Stories compared to its sale of regular advertisements.

    She says later:

    Plaintiffs assert that they have a tangible property interest in their personal endorsement of Facebook advertisers products to their Facebook Friends, and that Facebook has been unlawfully profiting from the nonconsensual exploitation of Plaintiffs statutory right of publicity. Thus, in the same way that celebrities suffer economic harm when their likeness is misappropriated for anothers commercial gain without compensation, Plaintiffs allege that they have been injured by Facebooks failure to compensate them for the use of their personal endorsements because [i]n essence, Plaintiffs are celebritiesto their friends.

    Clearly, Judge Koh is making a tricky intellectual move, and I bet it's going to make some privacy advocates unhappy. There is unquestionably a street value to data about a person to improve the marketing to that person, just as there is unquestionably commercial value in gaining an endorsement from a consumer. It's awkward to recognize one value and not the other. (Of course, in many of the precedent cases, there was only the possibility of data leakage; there wasn't actually a showing that any marketer had bought the leaked data for commercial reuse).

    However, Judge Koh's fancy footwork rips open only a very small hole in the Article III jurisprudence. Her exception only applies where there's a statutory publicity rights claim, and only when the defendant made a commercially-motivated endorsement. I'm sure we'll see plaintiffs advance claims to take advantage of this ruling, but few plaintiffs will be able to style their claims accordingly.

    In another tricky intellectual move, Judge Koh distinguishes Cohen v. Facebook, which dismissed a publicity rights claim based on Facebook's "Friend Finder" service, because this case showed a more direct connection between the friend's endorsement and the commercial value derived by Facebook. She also implies the lawyers did a better job here than in Cohen. I didn't fully understand this distinction other than Judge Koh's desire to reach a different result without disturbing the Cohen precedent.

    47 USC 230

    Facebook's 230 defense is tricky. First, it seeks to invoke the defense against a publicity rights claim, which the 9th Circuit said was possible in Perfect 10 v. ccBill in a controversial statutory reading that has been rejected by every other court outside the Ninth Circuit. Judge Koh doesn't touch that issue.

    Second, Facebook seeks 230 protection for the ad copy it created automatically. The ad is based on a user action, the "Like," plus various pieces of user content, but Facebook assembles it all into a package that the user never sees, blesses or necessarily even wants. We've had some other cases upholding 230 when a service provider is so intimately involved with creating the final content, such as the Carafano case, but Facebook is clearly playing at the edge of the statutory immunity.

    Judge Koh rules that Facebook is over that line and doesn't get the immunity. Unfortunately, she does so by saying that Facebook is partially the information content provider of the ads in question. She references the dispositive allegations:

    Plaintiffs allege that Facebook creates content by deceptively mistranslating members actions, such as clicking on a Like button on a companys page, into the words Plaintiff likes [Brand], and further combining that text with Plaintiffs photograph, the companys logo, and the label Sponsored Story. ... Plaintiffs allege that they themselves have no control over whether to post a particular companys name or logo, and that Facebook maintains sole control over whether to display a Sponsored Story at all.

    Personally, I'd be much more sympathetic to Facebook's position if users had the specific ability to "like" a business page without simultaneously authorizing the Sponsored Story. Because Facebook's controls are insufficiently granular, Facebook automatically interprets a "like" as both a statement of user attitudes and as a green light to create the Sponsored Story. In contrast, imagine that when a user "liked" a business page, Facebook prepared the ad copy for the Sponsored Story, presented it to the user, and asked the user if the user wanted to publish the ad copy to his/her friends. At this point, I would feel much more strongly that the ad copy really was the user's words. Naturally, Facebook doesn't give users this level of control over the words being put into their mouths.

    On the other hand, consider an alternative example where a website both publishes UGC on its site and then syndicates the content to third party sites. It's my position that the website gets 230 for both acts of publication, even if the user never expressly green-lighted the syndication (so long as the user-to-website license permitted the syndication). See, e.g., Prickett v. infoUSA. Based on Judge Koh's explication, I'm not exactly sure why Facebook crossed the 230 line while some of these other situations probably don't.

    Facebook responded that its activities didn't make it a content provider but just represented traditional editorial functions. The court rejects the argument, citing this allegation:

    Plaintiffs allege not only that Facebook rearranged text and images provided by members, but moreover that by grouping such content in a particular way with third-party logos, Facebook transformed the character of Plaintiffs words, photographs, and actions into a commercial endorsement to which they did not consent.

    In the context of this case, I see her point. Sadly, the opinion's wording will give false hope to a slew of plaintiffs who will argue that the website's presentation of third party content constituted some type of unauthorized endorsement. It will take a few cases to burst the plaintiffs' bubbles about a new exception to 230.

    The Statutory Publicity Rights Claim (CA Civil Code 3344)

    Facebook took a few cracks at the claim, all of which were unsuccessful:

    Newsworthiness. The publicity rights statute does not restrict using someone's personality "in connection with any news." This is a backdoor First Amendment defense, as what constitutes news tracks First Amendment jurisprudence on "matters of public interest." This defense seemed like a hail-mary for Facebook--a user "liking" a page is clearly "new" information to the marketplace, but it's not "news" in either the traditional or First Amendment sense. The court seems unimpressed, saying that even if a user "liking" a commercial product is news to that user's social network, using that information commercially drops out of the exception. I wasn't persuaded by the judge's distinction here, but then again Facebook's argument about what constituted "news" was obviously tendentious.

    I was a little disappointed that Judge Koh sidestepped some interesting lurking issues about what is "news" in the modern environment, where all of us are publishers to our local communities and we as publishers can have significant clout in a small community. Some academic literature in the 1990s discussed these issues in the Internet context, but it might be worth revisiting as a paper topic. Judge Koh also sidestepped the intellectually interesting issue of whether opinions about marketplace goods are "newsworthy," something that I strongly believe to be the case in the context of anti-SLAPP laws.

    Consent. Facebook argued that users consented to Sponsored Stories as part of its terms of use. The plaintiffs retorted that Sponsored Stories didn't exist when they signed up, so they couldn't have consented to it. The court says there's a factual dispute which prevents a motion to dismiss.

    Injury. Facebook argued that non-celebrities have to show economic injury as part of their 3344 prima facie case. The court rejects this distinction, saying "[i]n a society dominated by reality television shows, YouTube, Twitter, and online social networking sites, the distinction between a celebrity and a non-celebrity seems to be an increasingly arbitrary one." Furthermore, the plaintiffs did allege injury by showing that their endorsements were valuable to Facebook, which helps distinguish this case from the Cohen "Friend Finder" precedent. I liked this quote:

    While traditionally, advertisers had little incentive to exploit a non-celebritys likeness because such endorsement would carry little weight in the economy at large, Plaintiffs allegations suggest that advertisers ability to conduct targeted marketing has now made friend endorsements a valuable marketing tool, just as celebrity endorsements have always been so considered.

    For more on this point, see my Online Word of Mouth paper.

    Unfair Competition Law (UCL)

    Normally, we'd expect the UCL claim to be tossed because the plaintiffs can't make the required showing that they lost "money or property." Numerous Internet privacy cases have reached that conclusion. Judge Koh makes the same intellectual move she did with Article III standing, saying that publicity rights are different than other privacy torts. She says: "[t]o the extent Plaintiffs allege they can prove that their endorsement of commercial products to their Facebook Friends has concrete, quantifiable value for which they are entitled to compensation, the Court finds that Plaintiffs have properly alleged loss of money or property for purposes of establishing standing under the UCL." I wonder if plaintiffs can make that showing because there's no existing market for consumer-to-consumer endorsements, but it's enough to survive the motion to dismiss. In particular, she says California's statutory damages for publicity rights violations aren't enough to demonstrate the value of the endorsements.

    Judge Koh also concludes that plaintiffs properly alleged that Facebook's activities were unlawful, unfair and fraudulent (in the latter case, because Facebook allegedly overclaimed users' abilities to opt-out of Sponsored Stories).

    Unjust Enrichment

    Recent caselaw makes it even clearer that there's no separate cause of action for unjust enrichment; instead, it's just a synonym for restitution. As a result, the court tosses this claim.


    This is not a good ruling for Facebook, but I can't really feel too sorry for it. Facebook has been playing fast-and-loose with the law in many different contexts (see, e.g., its FTC bust), and Sponsored Stories is no different. Before rolling it out, Facebook surely knew that the Sponsored Stories offering was on murky legal ground. It can't be surprised that it didn't get an easy dismissal.

    Even so, if it gets that far, Facebook may yet win this case. Judge Koh has made it clear that she's a tough customer, but Facebook has plenty of power to its remaining arguments. Nevertheless, I'm reasonably confident it won't get that far. Given the importance of maximizing ad revenues and its desire to clean up legal issues in advance of an IPO, it seems more likely that Facebook will cut a deal with plaintiffs' counsel. I imagine Facebook might try to do a settlement like the Facebook Beacon settlement that results in minimal restrictive covenants, a chunk of money into the lawyers' hands, and a chunk of money that doesn't get into users' hands but instead goes into something like Facebook's privacy foundation.

    Read More | 5 Comments | Leave a Comment..

    Posted on Techdirt - 9 December 2011 @ 7:39pm

    The Good And The Bad Of The New OPEN Bill From Wyden And Issa

    from the some-good,-some-bad dept

    Sen. Wyden and Rep. Issa have released a draft of OPEN: Online Protection & ENforcement of Digital Trade Act, intended as an alternative to SOPA/PROTECT-IP. See my prior posts opposing SOPA and linkwrapping the discussion. Unlike SOPA's disgustingly blatant rent-seeking, which was such an over-the-top abuse of the legislative process that it did not (and could not) support a principled or even intelligent conversations about it, OPEN provides a useful starting point for a sensible conversation that could actually lead to acceptable compromises. For that reason alone, I think Congress should immediately stop all work on SOPA/PROTECT-IP and redirect that energy towards vetting this proposal. Having said that, for reasons I'll explain in a moment, I continue to believe the assumptions underlying SOPA/PROTECT-IP and OPEN are misguided, meaning that forging a compromise from OPEN's more sensible proposal may be tricky.

    Before I get further into substance, two process notes:

    First, SOPA was the product of rent-seekers who were talking only amongst themselves and legislators tethered to their campaign contributions. The drafting process was disturbingly closed-door and exclusionary, exactly the kind we wish didn't take place in our representative democracy. In contrast, the OPEN sponsors want to have a dialogue about their ideas. In support of that, they have posted the draft to a website that allows comments and discussion. This is the way our democracy SHOULD work. Why is such an open process the exception instead of the rule?

    Second, OPEN is a comparatively svelte 18 pages focused mostly on one core concept, compared to SOPA's 78 page monstrosity that advanced about a dozen different substantive proposals. I can't tell you the number of times I've seen very smart people stymied to keep all of SOPA's moving parts separate, and the failure to do so meant that they were conflating different parts of the statute in ways that prevented productive discussion. (Just two examples: the Colbert Report, where Zittrain mostly focused on SOPA's felony streaming provision while his counterpart was mostly talking about the cutoff provisions; and Business Insider's infographic where the felony streaming sanction was presented as a remedy to the cutoff provisions.) By reducing the number of topics at issue, OPEN substantially reduces the chance that policy discussants will simply talk past each other.

    An Overview

    The law contemplates that rightsowners can file a petition against rogue websites at the ITC, an independent federal agency best known for its adjudication of certain patent disputes. In response to the rightsowner's petition, the ITC will conduct an administrative adjudication. If the ITC determines that the website is a rogue website, then (1) the website is required to cease its conduct (not sure how enforceable that is), (2) the site also will be subject to any other unspecified consequences following from its determination as a rogue actor, and (3) most importantly, the rightsowner can take the ITC determination to payment service providers (PSPs) and ad networks and have them cut off the flow of money to the rogue website. The PSPs and ad networks would be protected by several immunities for trying to comply with the orders or their other efforts to protect the public.

    This makes OPEN similar to SOPA in that it seeks to cut off funds flowing to rogue actors. However, among other key differences, PSPs and ad networks have no legal obligations until the ITC makes a ruling. In contrast, SOPA imposed cutoff obligations on PSPs and ad networks based merely on rightsowners' unsubstantiated assertions.

    What's Good

    Substantively, some of the things I liked about OPEN:

    * it situates the discussion about "rogue websites" in foreign trade policy. This fixes SOPA's overinclusive application to both domestic and foreign actors. However, if we really think rogue websites are a transborder enforcement problem, there are many other trade policy solutions that might be better options to considerthe most obvious being transborder enforcement coordination like the FTC does with its foreign counterparts.

    * OPEN doesn't touch the domain name system or search engines. SOPA had the potential to destroy the DNS and to jeopardize search engine functioning. OPEN sidesteps both pitfalls.

    * OPEN builds in some due process before any formal legal obligations attach. As we've recently seen, due process is actually quite important, and we suffer from its absence. I say "some" due process because I'm not sure how much due process will attach in practice. For example, I have some concerns about the notice provision--not every targeted website will receive notice of the ITC investigation. However, I did like that any website the ITC labels as rogue can correct any identified problems, reapproach the ITC and ask it to remove the "rogue" determination.

    * the definition of rogue website is tightened up substantially. It requires three elements:
    a) a "non-domestic domain name," which requires that the registry, registrar and registrant all have to be located outside the US (I'm not sure what "located" means in this context). Venkat asked me what happens to a .com registered with a foreign registrar; I believe OPEN does not apply to this domain name.
    b) conducting business in the US; and
    c) "has only limited purpose or use other than engaging in infringing activity and whose owner or operator primarily uses the site to willfully engage in infringing activity."

    The last element, in particular, is quite restrictive by requiring willful infringement. The meaning of the word "willful" is notoriously murky (see, e.g., the multitudinous Supreme Court cases over the word), so the statute would be improved by using a more detailed synonym. No matter what, though, willful is a high scienter level that should easily exclude most legitimate players. The statute further expressly excludes any sites that:

    - follow good notice-and-takedown procedures
    - qualify for 17 USC 512 (the DMCA online safe harbors) [this means that the statute sits next to 512 instead of rendering 512 moot like SOPA threatened to do], or
    - distribute "copies that were made without infringing a copyright or trademark." I'm not 100% sure what this means. It apparently excludes websites reselling goods covered by the First Sale doctrine. I presume that the exclusion includes sites that sell legitimate knock-off goods, such as replicas of goods that aren't protected by copyrights or trademarks.

    * if a PSP or ad network fails to comply with an ITC order, the only consequence is that the DOJ can seek injunctive relief. Rightsowners do not have a private cause of action in those cases. As discussed below, this doesn't eliminate all PSP/ad network exposure to rightsowners, but rightsowners can't introduce evidence of ITC orders in any civil suits they bring against PSPs or ad networks.

    * on the trademark side, it expressly limits its applicability to counterfeiting (although there is an erroneous cross-reference in the draft). Presumably, dilution or garden-variety trademark infringement disputes don't qualify under the statute.

    What's Not Good

    Substantively, some of the things I don't like about OPEN:

    * OPEN still contemplates reestablishing a Fortress USA. Fortress USA marginally makes sense regarding the shipment of physical goods across geographic borders. It makes zero sense for digital bits zinging around the borderless network.

    * in particular, because OPEN would burden only US-governed PSPs and ad networks, it may drive websitesincluding legitimate websites who want to reduce their risk of being mistargetedto shift their business to foreign-based PSPs and ad networks. If lots of businesses make a switch based on these concerns, OPEN could counterproductively result in net financial losses for the US economy.

    * similarly, foreign websites can opt-out entirely of the ITC process by consenting to US judicial jurisdiction. I like the idea of an opt-out, but imagine if other countries offered the same quid-pro-quo of allowing US websites to opt-out of some nasty foreign process so long as the websites consent to jurisdiction in their countries. I think we'd be outraged and insulted; which is how I would expect foreign countries to view this quid-pro-quo. Cf. Venkat's recent post on Facebook v. Faceporn. Then again, other countries might think it's a pretty good idea, leading to a proliferation of transborder quid-pro-quo jurisdictional offers.

    * designating the ITC to conduct the investigations is a little odd. First, the ITC is an administrative agency, not a federal court. I don't fully understand all of the implications of administrative vs. judicial review, but I believe there are substantial procedural differences that could lead to important substantive differences. Second, the ITC has been gamed in the patent world (see, e.g., my colleague Colleen Chien's research on the ITC explaining how the ITC hears many US company vs. US company disputes), so I fear similar gaming will emerge. For example, a rightsowner chasing a rogue website could simultaneously pursue a domestic court action, a foreign court action and an ITC proceeding. How would these types of parallel proceedings play out in practice? We're still trying to resolve the parallel proceeding problems in patents.

    * like SOPA, the bill covers copyright infringement, trademark infringement *and* 1201 circumvention. I don't understand why the circumvention issue is getting equal billing or how often transborder circumventions are a real problem. Seeing how 1201 circumvention lawsuits have devolved into anti-competitive enforcements, picking up the circumvention piece could increase the risk of competitive misuse of the statute.

    * like SOPA, the definitions are vague. Consider, for example, the definition of Internet advertising service:

    The term Internet advertising service means a service that serves an online advertisement in viewable form for any period of time on an Internet site.

    Hmm...what does that mean? Notice that the definition doesn't directly distinguish between third-party ad networks and sites that sell their own ads. I think in practice sites that sell their own ads drop out of the statute, so one possible implication is that more sites will ramp up their own ad sales. (This is doubtful, but just throwing the possibility out there.) I think the focus on "viewable" is interesting; are audio-only ads excluded? And what does it mean to "serve" content? This contemplates a specific technological interaction that I don't fully understand today and will almost certainly evolve over time.

    Why I'm Not Enthusiastic About OPEN

    Even though OPEN is worth discussing intelligently, unlike SOPA, I believe it's based on two underlying assumptions that aren't fixable.

    First, like SOPA, OPEN assumes there is a problem with foreign rogue websites that needs to be solved. I'm not saying there isn't, but the policy discussions have been startlingly devoid of reliable and credible facts demonstrating the nature and scope of the problem.

    Instead, the evidence in support of a rogue website "problem" typically consists of two main threads: (a) people are dying from counterfeit drugs, and (b) bad guys are "stealing" our stuff. With respect to the former, I've never seen anything more than ad hoc assertion; but if there's a real problem, counterfeit drugs can be fixed with a highly targeted solution. With respect to the latter, it's hard to give those arguments much credit. After all, all rightsowners' arguments are inherently self-interested: it's in their financial interest to say that they would like to make more money than they are making. It's also in their interest to bemoan broad sectoral changes in the economy as evidence that someone is capturing money they think they are entitled to (and to use rent-seeking to thwart those broad sectoral changes). More importantly, there is lots of evidence that a lot of rightsowners are making a lot of money today, both via the Internet and more generally. So it's hard to break out the quantity of actual economic losses that rightsowners are truly suffering when those claims are intermingled with rightsowners' general rent-seeking efforts.

    Therefore, until the rightsowners offer us more than the trumped-up BS already-discredited statistics, I'm still not clear on the problem, how bad it is, how any legislative solution would remediate that problem, and if the collateral consequences of the effort to remediate the problem are greater or less than the problem itself. OPEN does nothing to fill the void of supporting foundational evidence of the problem, so it's hard for me to be enthusiastic about its solution.

    Second, and more importantly, attacking the money supply to supposed bad actors remains too blunt an instrument. I may be truly on my own on this point, as many people I respect--including, notably, Rep. Lofgren--are prepared to embrace the policy solution of cutting off money flows. However, by embracing an attack on the movement of money, OPEN replicates one of SOPA's sins. If a player is engaged in legitimate and illegitimate activity and its money supply is cut off, both activities go down the tubes. In contrast, one of the positive aspects of 17 USC 512(c) and (d) is that they require the copyright owner to identify infringing items and target only those items. Giving rightsowners a remedy that would affect an entire site for only some items on the site goes too far.

    The OPEN bill tries hard to minimize overbreadth by narrowly defining the targeted websites. Perhaps this definition is narrow enough that there won't be much collateral damage. However, in practice, regulating money flows nevertheless could have pernicious effects in the field. A PSP or ad network drawn into an ITC proceeding frequently will "voluntarily" choose to toss the targeted website before the ITC proceeding reaches its conclusioneven if the ITC proceeding would have rejected the challenge. Furthermore, rightsowners still will send cutoff notices to PSPs/ad networks without filing any ITC petition, and the PSPs/ad networks will often honor them as a way of preempting an ITC proceeding.

    What this teaches me (in combination with the Elsevier v. Chitika case) is that PSPs and ad networks need robust statutory immunities which are not based on a notice-and-takedown scheme. On the trademark side, the need for an immunity became clear after the sloppy language in Gucci v. Frontline. On the copyright side, 512 doesn't cover PSPs and ad networks, probably because in a million years the safe harbor drafters never thought PSPs and ad networks would be liable for third party infringing activity in the first place. Now that we've seen copyright law and trademark law creep much further than we could have imagined in 1998, we should plug this liability hole completely. If OPEN proceeds, it should have a broad-based immunity for PSPs and ad networks with the idea that rightsowners are getting a specific remedy against them in the new law.

    While OPEN can't really be fixed to resolve my two structural concerns, my hope is that the discussion about OPEN will force rightsowners to provide *credible* evidence of harms that they or consumers are suffering (no more self-serving hype, please), and that such evidence will force us to think carefully about how "rifle shot" solutions (as opposed to shotgun solutions) can ameliorate those harms. If we have a discourse that even slightly resembles this ideal, then OPEN will be successful no matter what final outcome we reach.

    75 Comments | Leave a Comment..

    Posted on Techdirt - 7 December 2011 @ 12:09am

    Court Recognizes Daily Groupon Deal Hunters Aren't Likely To Be Confused By Groupion's Enterprise Software

    from the but-i-want-half-off-my-crm! dept

    Groupion provides CRM software as a service (SaaS). Groupon distributes "deal of the day" offers that are typically unprofitable for advertisers and often have the extra "benefit" of causing the advertisers to get trashed on Yelp. Groupion sued Groupon for trademark infringement. I previously blogged on the complaint.

    The court denies Groupion's various motions. The court runs through a typical multi-factor likelihood of consumer confusion analysis:

    • mark similarity. Calling this factor "critical," the court concludes that consumers can keep the marks separate. The visual depictions of the logos are different, Groupion has one more syllable, and the words are portmaneaus from different inspirations ("coupon" + "group" vs. "groupware" + "companion").
    • product similarity. "The parties' products are used for different functions and purposes, and are purchased by different classes of consumers."
    • marketing channels. The Internet's commonality is discounted (cite to Network Automation), and the rest of the parties' channels are disparate as you would expect when one company is B2B and the other B2C.
    • mark strength. The court says "Groupion" is weak because others are using similar marks in its field and Groupion didn't show evidence of the mark's commercial strength.
    • intent. Groupion asserted that Groupon must have selected the mark knowing about it, but Groupion didn't provide any evidence to that effect. Groupon's failure to stop after Groupion's C&D was irrelevant.
    • evidence of actual confusion. None of Groupion's evidence "demonstrates instances of actual confusion by its customers regarding the source of its products."
    • likelihood of expansion. Groupon bought a mobile apps business, but the court says it will use that to distribute deals, not for intracompany groupware.
    • purchaser care. B2B software buyers are careful, as are advertisers.
    Thus, the factors point against Groupion's likelihood of success, so the court denies the preliminary injunction, Groupion's motion for summary judgment and motion to cancel Groupion's registered mark. I assume Groupon will view this as an invitation to bring its own summary judgment motion and kill this case. The judge's tone was unmistakable.

    11 Comments | Leave a Comment..

    This site, like most other sites on the web, uses cookies. For more information, see our privacy policy. Got it