ExTwitter Unfortunately Loses Round One In Challenging Problematic Content Moderation Law

Content Moderation

from the well,-that's-unfortunate dept

Back in September we praised Elon Musk for deciding to challenge California’s new social media transparency law, AB 587. As we had discussed while the bill was being debated, while it’s framed as a transparency bill, it has all sorts of problems. It would (1) enable the California government officials (including local officials) to effectively put pressure on social media companies regarding how they moderate by enabling litigation for somehow failing to live up to a terms of service, (2) make it way more difficult for social media companies to deal with bad actors by limiting how often they can change their terms of service, and (3) hand bad and malicious actors a road map for being able to claim they’re respecting the rules, while clearly abusing them.

Yet, the largest social media companies (including Meta and Google) apparently are happy with the law, because they know it creates another moat for themselves. They can deal with the compliance requirements of the law, but they know that smaller competitors cannot. And, because of that, it wasn’t clear if anyone would actually challenge the law.

A few Twitter users sued last year, but with a very silly lawyer, and had the case thrown out because none of the plaintiffs had standing. But in the fall, ExTwitter filed suit to block the law from going into effect, using esteemed 1st Amendment lawyer Floyd Abrams (though, Abrams has had a series of really bad takes on the 1st Amendment and tech over the past decade or so).

The complaint still seemed solid, and Elon deserved kudos for standing up for the 1st Amendment here, especially given the larger tech companies’ unwillingness to challenge the law.

Unfortunately, though, the initial part of the lawsuit — seeking a preliminary injunction barring the law to go into effect — has failed. Judge William Shubb has sided with California against ExTwitter, saying that Elon’s company has failed to show a likelihood of success in the case.

The ruling relies heavily on a near total misreading of the Zauderer case, regarding whether or not compelled commercial speech was allowed under the 1st Amendment. As we discussed with Professor Eric Goldman a while back, reading Zauderer, you see that the case was ruled on narrow grounds, saying you could mandate transparency if it was about the text in advertisements, required disclosure of purely factual information, the information disclosed would be uncontroversial, and required the disclosure to be about the terms of an advertiser’s service. If all those conditions are met, the law might still be found unconstitutional if the disclosure requirements are not related to preventing consumer deception or if the disclosure requirements are unduly burdensome.

As professor Goldman has compellingly argued, laws requiring social media companies reveal to government officials their moderation policies meet basically none of the Zauderer conditions. It’s not about advertising. It’s not purely factual information. The disclosures can be extremely controversial. The disclosures are not about any advertiser’s services. And, on top of that, it has nothing to do with preventing consumer deception and the requirements can be unduly burdensome.

A New York Court threw out a similar law, recognizing that Zauderer shouldn’t be stretched this far.

Unfortunately, Shubb goes the other way, and argues that Zauderer makes this kind of mandatory disclosure compatible with the 1st Amendment. He does so by rewriting the Zauderer test, leaving out some of the important conditions, and then mis-applying the test:

Considered as such, the terms of service requirement appears to satisfy the test set forth by the Supreme Court in Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626 (1985), for determining whether governmentally compelled commercial disclosure is constitutionally permissible under the First Amendment. The information required to be contained in the terms of service appears to be (1) “purely factual and uncontroversial,” (2) “not unjustified or unduly burdensome,” and (3) “reasonably related to a substantial government interest.”

The court admits that the compelled speech here is different, but seems to think it’s okay, citing both the 5th and 11th Circuits in the NetChoice cases (who both also applied the Zauderer test incorrectly — which is why we pointed out this part of the otherwise strong 11th Circuit decision was going to be a problem):

The reports to the Attorney General compelled by AB 587 do not so easily fit the traditional definition of commercial speech, however. The compelled disclosures are not advertisements, and social media companies have no particular economic motivation to provide them. Nevertheless, the Fifth and Eleventh Circuits recently applied Zauderer in analyzing the constitutionality of strikingly similar statutory provisions requiring social media companies to disclose information going well beyond what is typically considered “terms of service.”

Even so, this application of the facts to the misconstrued Zauderer test… just seems wrong?

Following the lead of the Fifth and Eleventh Circuits, and applying Zauderer to AB 587’s reporting requirement as well, the court concludes that the Attorney General has met his burden of establishing that that the reporting requirement also satisfies Zauderer. The reports required by AB 587 are purely factual. The reporting requirement merely requires social media companies to identify their existing content moderation policies, if any, related to the specified categories. See Cal. Bus. & Prof. Code § 22677. The statistics required if a company does choose to utilize the listed categories are factual, as they constitute objective data concerning the company’s actions. The required disclosures are also uncontroversial. The mere fact that the reports may be “tied in some way to a controversial issue” does not make the reports themselves controversial.

But… that’s not even remotely accurate on multiple accounts. It is not “purely factual information,” that is required to be disclosed. The disclosure is about the highly subjective and constantly changing processes by which social media sites choose to moderate. Beyond covering way more than merely factual information, it’s also extraordinarily controversial.

And that’s not just because they’re often tied to controversial issues, but rather because users of social media are constantly “rules litigating” moderation decisions, and insisting that websites should or should not moderate in certain ways. The entire point of this law is to try to pressure websites to moderate in a certain way (which alone should show the Constitutional infirmities in the law). In this case, it’s California trying to force websites to remove “hate speech” by demanding they reveal their hate speech policies.

Now, assuming most of you don’t like hate speech, you might not see this as all that controversial, but if that’s allowed, what’s to stop other states from requiring the same thing regarding how companies deal with other issues, like LGBTQ content. Or criticism of the police.

But, the court here insists that this is all uncontroversial.

And worse, it ignores that the Zauderer test is limited only to issues of consumer deception.

The California bill has fuck all to do with consumer deception. It is entirely about pressuring websites in how they moderate.

Also, Shubb shrugs off the idea that this law might be unduly burdensome:

While the reporting requirement does appear to place a substantial compliance burden on social medial companies, it does not appear that the requirement is unjustified or unduly burdensome within the context of First Amendment law.

The Court also (again, incorrectly in my opinion) rejects ExTwitter’s reasonable argument that Section 230 pre-empts this. Section 230 explicitly exempts any state law that seeks to limit a website’s independence in making moderation decisions, and thus this law should be pre-empted as such. Not so, says the court:

AB 587 is not preempted. Plaintiff argues that “[i]f X Corp. takes actions in good faith to moderate content that is ‘obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable,’ without making the disclosures required by AB 587, it will be subject to liability,” thereby contravening section 230. (Pl.’s Mem. (Docket No. 20) at 72.) This interpretation is unsupported by the plain language of the statute. AB 587 only contemplates liability for failing to make the required disclosures about a company’s terms of service and statistics about content moderation activities, or materially omitting or misrepresenting the required information. See Cal. Bus. & Prof. Code § 22678(2). It does not provide for any potential liability stemming from a company’s content moderation activities per se. The law therefore is not inconsistent with section 230(c) and does not interfere with companies’ ability to “self-regulate offensive third party content without fear of liability.” See Doe, 824 F.3d at 852. Accordingly, section 230 does not preempt AB 587.

Again, this strikes me as fundamentally wrong. The whole point of the law is to force websites to moderate in a certain way, and to limit how they can moderate in many scenarios, thus creating liability for moderation decisions regarding whether or not those decisions match the policies disclosed to government officials under the law. That seems squarely within the pre-emption provisions of Section 230.

This is a disappointing ruling, though it is only at stage one in this case. One hopes that Elon will appeal the decision and hopefully the 9th Circuit has a better take on the matter.

Indeed, I’d almost hope that this case were one that makes it to the Supreme Court, given the makeup of the Justices on the Supreme Court today and the (false, but whatever) belief that Elon has enabled “more free speech” on ExTwitter. It seems like this might be a case where the conservative Justices might finally understand why these kinds of transparency laws are problematic, by seeing how California is using them (as opposed to the Florida and Texas laws it’s reviewing currently, where that wing of the Supreme Court is more likely willing to side with those states and their goals).

Filed Under: 1st amendment, ab 587, california, elon musk, mandated transparency, rob bonta, section 230, terms of service, transparency, william shubb, zauderer
Companies: twitter, x