Elon Musk Files Really Strong 1st Amendment Challenge To California’s Terrible Social Media ‘Transparency’ Law
from the good-job-elon! dept
Hey, Techdirt haters: hold onto your hats, because I’m going to praise Elon Musk for doing the right thing, even though many of you insist that my complaints about him are motivated by personal dislike. But, as I’ve noted repeatedly, I’m happy to highlight when he does the right thing, such as here where he is (perhaps surprisingly) challenging a terrible internet law that the bigger internet companies refuse to challenge (because it helps them), and bringing in a big time 1st Amendment lawyer to do so. And, importantly, the challenge seems really well done.
Over the last few years we’ve been writing about a long list of awful internet laws in California. Last year, the one that got the most attention was AB 2273, or the Age Appropriate Design Code, which we’re still hoping a federal court will toss out as unconstitutional in the near future. However, there was another bill we wrote about that got a bit less attention: AB 587 which was pitched as a “transparency” bill.
As we’ve said many times in the past, more transparency is a very good and important goal, and one that people should demand of the companies whose products and services they use. But mandated transparency creates all sorts of problems. AB 587 creates oh so many problems. So, of course, Governor Gavin Newsom signed it into law.
For what it’s worth, I’ve heard from multiple sources that the big internet companies love AB 587. It creates another moat for them that harms smaller companies. It allows them to tell lawmakers that they’re happy to embrace some laws that those big companies can handle, even if they’re problematic.
But it’s a bad law. It limits the ability of companies to respond and adjust tactics against bad actors in real time. It assumes (incorrectly) that content moderation is a static thing. It also requires that companies have policies in place to deal with 1st Amendment protected content, even though the government can’t legislate that.
Worse, the transparency requirements will be weaponized by malicious actors, using them to force websites to host content and to publicly litigate moderation decisions, by claiming they don’t comply with the posted terms. While the law has no private right of action, it does enable any local DA or county lawyer to bring a lawsuit, and you can imagine how many local DAs looking to get headlines will have a field day going after “big tech.”
Of course, since the big internet companies are fine with the law, I wasn’t sure if anyone would challenge it. Earlier this year I was surprised (and a bit worried) when Minds Inc., Tim Pool, and the Babylon Bee… all challenged the law (later they added the National Religious Broadcasters to try to have at least one plaintiff who might meet the qualifications for standing). I’d been meaning to write about that case over the last few months, but never got around to it. I was worried, because the lawyer on their case was James Lawrence from Envisage Law, who has filed some truly ridiculously dumb lawsuits, claiming that content moderation violates the 1st Amendment. And the complaint in that case… was not well done.
So, while I was happy someone was challenging the law, it was frustrating to see that it was being done with nonsense-peddling plaintiffs and a lawyer with a pedigree to match.
And, indeed, that lawsuit was just dismissed a few weeks ago, because Lawrence couldn’t explain to the court how any of his clients actually had standing under the law, since there’s basically no chance 587 will impact any of them:
In a pre-enforcement action, which this case undoubtedly is, “Plaintiffs must demonstrate a realistic danger of sustaining a direct injury as a result of the statute’s operation or enforcement.” Libertarian Party of Los Angeles Cnty. v. Bowen, 709 F.3d 867, 870 (9th Cir. 2013) (quoting Babbitt v. United Farm Workers Nat’l Union, 442 U.S. 289, 298 (1979)) (internal quotation marks omitted) (emphasis added). A plaintiff may satisfy the “concreteness” requirement by alleging an injury that actually exists or by alleging a risk of real harm. Spokeo, Inc., 578 U.S. at 340. In First Amendment cases, the Supreme Court has required plaintiffs to demonstrate “a claim of specific present objective harm or a threat of specific future harm.” Laird v. Tatum, 408 U.S. 1, 14 (1972).
Plaintiffs’ allegations do not satisfy these important foundational requirements for justiciability. Plaintiffs’ first theory of standing—that AB 587 will make it “more likely” Plaintiffs’ content will be moderated or censored—is simply conjecture and is unsupported by any allegations in the FAC that Plaintiffs have suffered direct injuries in the present or that they face a “realistic danger” of specific harm in the future. See Libertarian Party of Los Angeles Cnty, 709 F.3d at 870. Nor do they posit any coherent theory of how any censorship or harm they may suffer could be “fairly traceable” to AB 587, or how it is “real, immediate, and direct.”
The judge noted that the chain of events the complaint laid out is not nearly complete enough to show a likelihood of harm. As Eric Goldman explains, the steps that the plaintiffs argued could lead to 587 pressuring companies to take down their content are “absolutely right,” but the court brushes it off as containing too many “logical leaps.”
And thus… case dismissed. While the court offers the plaintiffs a chance to amend, it’s pretty clear that these plaintiffs are not particularly strong ones to challenge this law.
However, one thing that the Babylon Bee (especially) and to a lesser extent Minds and Tim Pool have in common is that Elon Musk seems to pay attention to them (for better or for worse). And Elon Musk now (regretfully) owns a platform that actually is very clearly subject to 587’s transparency mandates.
And thus… Elon Musk’s exTwitter is actually challenging the law. And they’re using one of the most famous 1st Amendment lawyers, Floyd Abrams, to do so. That’s way better than “Envisage Law.” Abrams isn’t always great on the 1st Amendment and the internet: back in the SOPA days he argued (for his long term client the MPAA) that even though SOPA would censor protected speech, it was acceptable collateral damage. And, more recently, he’s defended Clearview’s questionable selling of scraped facial data to cops on 1st Amendment grounds. But the complaint here is really, really strong. This is a good thing.
So, kudos to Elon Musk for actually challenging this terrible law (and a big boo to all the other big social media sites for failing to do so). I’m glad some social media company is actually challenging the law and doing so well.
As for the actual complaint, it tells the story clearly:
AB 587 violates the First Amendment of the United States Constitution and Article I, Section 2, of the California Constitution because it compels companies like X Corp. to engage in speech against their will, impermissibly interferes with the constitutionally-protected editorial judgments of companies such as X Corp., has both the purpose and likely effect of pressuring companies such as X Corp. to remove, demonetize, or deprioritize constitutionally-protected speech that the State deems undesirable or harmful, and places an unjustified and undue burden on social media companies such as X Corp.
The State of California touts AB 587 as a mere “transparency measure” under which certain social media companies must make their content moderation policies and statistics publicly available. See Press Release, Governor Newsom Signs Nation-Leading Social Media Transparency Measure (Sept. 13, 2022), https://www.gov.ca.gov/2022/09/13/governor-newsom-signs-nationleading-social-media-transparency-measure/. Yet, a review of the law’s purpose and likely effect — as evidenced by the legislative history and statements from AB 587’s author, sponsors, and supporters — demonstrates otherwise. As made clear by both the legislative history and public court submissions from the Attorney General in defending the law, the true intent of AB 587 is to pressure social media platforms to “eliminate” certain constitutionally-protected content viewed by the State as problematic. Ex. 1 (Cal. Assemb. Comm. on Judiciary Report, 2021–22 Sess. (AB 587), Apr. 27, 2021) at 4 (“if social media companies are forced to disclose what they do in this regard [i.e., how they moderate online content], it may pressure them to become better corporate citizens by doing more to eliminate hate speech and disinformation.”); Mot. to Dismiss at 15–16, Minds, Inc., et al. v. Bonta, No. 23-cv-2705 (ECF 23-1) (C.D. Cal. May 25, 2023) (“[T]he Legislature also considered that, by requiring greater transparency about platforms’ content-moderation rules and decisions, AB 587 may result in public pressure on social media companies to ‘become better corporate citizens by doing more to eliminate hate speech and disinformation’ on their platforms. . . . This, too, is a substantial state interest.”). AB 587 is, according to the law’s lead author, Assembly Member Jesse Gabriel, an “important first step in protecting our democracy from the dangerously divisive content that has become all too common on social media.” Ex. 1 at 4. The legislative record is crystal clear that one of the main purposes of AB 587 — if not the main purpose — is to pressure social media companies to eliminate or minimize content that the government has deemed objectionable.
As the complaint makes clear, this is compelled speech, and it includes compelled speech on fraught political issues:
The topics that AB 587 forces social media platforms to speak about against their will are highly controversial and politically charged. As the legislative history acknowledges, the categories of speech on which AB 587 focuses are those that are difficult to define because their boundaries are “often fraught with political bias.” Ex. 2 (Cal. Assemb. Comm. on Privacy and Consumer Protection Report, 2021–22 Sess. (AB 587), Apr. 22, 2021) at 4. And social media companies are frequently criticized, no matter what they do, by individuals on both sides of the political aisle, for their editorial decisions about speech that arguably falls into these ill-defined categories. The Assembly Reports from the Committee on Privacy and Consumer Protection describe the “complex dilemma” that social media companies increasingly find themselves in when trying to define and moderate the politically-charged and controversial categories of content that are the focus of AB 587:
As online social media become increasingly central to the public discourse, the companies responsible for managing social media platforms are faced with a complex dilemma regarding content moderation, i.e., how the platforms determine what content warrants disciplinary action such as removal of the item or banning of the user. In broad terms, there is a general public consensus that certain types of content, such as child pornography, depictions of graphic violence, emotional abuse, and threats of physical harm are undesirable, and should be mitigated on these platforms to the extent possible. Many other categories of information, however, such as hate speech, racism, extremism, misinformation, political interference, and harassment [i.e., the categories that are the focus of AB 587], are far more difficult to reliably define, and assignment of their boundaries is often fraught with political bias. In such cases, both action and inaction by these companies seems to be equally maligned: too much moderation and accusations of censorship and suppressed speech arise; too little, and the platform risks fostering a toxic, sometimes dangerous community
As the complaint rightly notes (and as we argued in our posts about the problems of this bill):
In other words, AB 587 seeks to force social media companies to provide the Attorney General and the public detailed information about how, if at all, they define and moderate the boundaries of the most controversial categories of content — i.e., those categories with no “general public consensus” about how to define or moderate them because their boundaries are “fraught with political bias” and are “difficult to reliably define” — and provide detailed information about what actions they have or have not taken in regulating those controversial categories, even though action or inaction “seems to be equally maligned” by members of the public, depending on their political viewpoint. Put another way, through AB 587, the State is compelling social media companies to take public positions on controversial and politically-charged issues. And, because X Corp. must take such positions on these topics as they are formulated by the State, X Corp. is being forced to adopt the State’s politically-charged terms, which is a form of compelled speech in and of itself.
AB 587 thus mandates X Corp. to speak about sensitive, controversial topics about which it does not wish to speak in the hopes of pressuring X Corp. to limit constitutionally-protected content on its platform that the State apparently finds objectionable or undesirable. This violates the free speech rights granted to X Corp. under the First Amendment to the United States Constitution and Article I, Section 2, of the California Constitution.
The 1st Amendment argument is a strong one:
The First Amendment unequivocally prohibits this kind of interference with a traditional publisher’s editorial judgment. For example, if a law were to require newspapers to promulgate Terms of Service that disclose (i) a time by which they will respond to requests to publish letters to the editor or opinion pieces; (ii) detailed disclosures about their criteria for publication and statistics about the bases for decisions regarding whether to publish letters to the editor or opinion pieces; and (iii) statistics about how many submissions were accepted and rejected on the ground that they contained “hate speech” or “misinformation,” it would undoubtedly violate the First Amendment because it would impermissibly interfere with the constitutionally-protected editorial judgment of newspapers. See, e.g., Herbert v. Lando, 441 U.S. 153, 174 (1979) (a law that “subjects the editorial process to private or official examination merely to satisfy curiosity or to serve some general end such as the public interest . . . would not survive constitutional scrutiny as the First Amendment is presently construed”); id. at 172 (concluding that “if inquiry into editorial conclusions threatens the suppression . . . of truthful information,” it raises First Amendment problems); Miami Herald Pub. Co. v. Tornillo, 418 U.S. 241, 258 (1974) (“The choice of material to go into a newspaper, and the decisions made as to limitations on the size and content of the paper, and treatment of public issues and public officials — whether fair or unfair — constitute the exercise of editorial control and judgment. It has yet to be demonstrated how governmental regulation of this crucial process can be exercised consistent with First Amendment guarantees of a free press as they have evolved to this time.”).
The complaint cites the Supreme Court’s excellent Halleck decision to highlight how social media still has editorial rights over content on its platform:
That X is a social media platform does not change the analysis.1 These core First Amendment principles prohibit the government from interfering with the right of private parties, like X Corp., to exercise “editorial control over speech and speakers on their properties or platforms.” Manhattan Cmty. Access Corp. v. Halleck, 139 S. Ct. 1921, 1932 (2019). And the U.S. Supreme Court has made clear that the right to choose whether to speak and how to tailor one’s speech is not “restricted to the press.” Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Boston, 515 U.S. 557, 574 (1995). This fundamental right applies equally to “business corporations generally” and “ordinary people engaged in unsophisticated expression as well as professional publishers.” Id. A private social media company’s editorial judgment about how to regulate content on its platform and what to say (or not to say) about its regulation of content is therefore fully protected under the First Amendment. Id.; see also U.S. Telecom Ass’n v. FCC, 855 F.3d 381, 435 (D.C. Cir. 2017) (Kavanaugh, J., dissenting from denial of rehearing en banc) (the government may not “tell Twitter or YouTube what videos to post” or “tell Facebook or Google what content to favor” any more than it may “tell The Washington Post or the Drudge Report what columns to carry”); Tornillo, 418 U.S. at 258 (editorial control and judgment protected from government regulation by First Amendment).
And (yes!) the complaint highlights the burdens on speech that 587 creates:
AB 587 also imposes tremendously burdensome requirements on social media companies, requiring them to keep records about potentially hundreds of millions of content moderation decisions made on a daily basis. Ex. 2 at 4. Worse yet, it threatens draconian financial penalties of up to $15,000 per violation per day if compliance is not made in “reasonable, good faith,” a term that the statute does not define and that gives the Attorney General nearly unfettered discretion to threaten to impose draconian fines if social media companies’ content moderation policies are not to the State’s liking. Given the broad enforcement powers granted to the Attorney General under Cal. Gov’t Code §§ 11180–81, AB 587 empowers the Attorney General to issue civil investigative demands to social media companies about their content moderation policies and practices to determine if they have complied with the statute in “reasonable, good faith.” This broad, unfettered discretion to scrutinize the editorial judgments of social media companies empowers the Attorney General to use his enforcement powers to pressure social media companies to regulate content in ways that the government wants — which is one of the stated purposes of the law.
Also (this is news!) the complaint notes that California Attorney General Rob Bonta has already threatened exTwitter over this law:
Concerns about the potential use of AB 587 in this manner are not speculative: Attorney General Bonta has already written to X Corp. (and other social media companies) threatening that “[t]he California Department of Justice will not hesitate to enforce [AB 587],” while in the same proverbial breath reminding the companies of their “responsibility” to combat what the Attorney General views as the “dissemination of disinformation that interferes with our electoral system.” Letter from Attorney General Robert Bonta to Twitter, Inc., et al., 4 (Nov. 3, 2022),
Finally, the complaint argues (like the challenges to other state laws) that 587 violates both the Dormant Commerce Clause (regulating interstate activity by a state) and Section 230’s preemption against state laws getting into content moderation.
The company is asking for a preliminary and permanent injunction against the law, saying it violates the 1st Amendment, the Dormant Commerce Clause, and preemption under Section 230, and it’s also asking for attorneys fees.
This is really good. Kudos to Elon Musk for (finally!) actually defending free speech, rather than his warped sense of free speech. This is a 1st Amendment challenge worth bringing, and Musk deserves credit for doing so.
Filed Under: 1st amendment, ab 587, california, content moderation, dormant commerce clause, elon musk, floyd abrams, gavin newsom, section 230, terms of service, transparency
Companies: minds, twitter, x