Analysis

In Hearing, Judge Seemed Torn on NetChoice’s Renewed First Amendment Attack on California’s Age-Appropriate Design Code

January 24, 2025 | Megan Iorio, Senior Counsel

In a hearing yesterday on NetChoice’s second motion for a preliminary injunction against California’s Age-Appropriate Design Code (CAADC), Judge Beth Labson Freeman appeared far more skeptical of the trade association’s First Amendment arguments than the first time she ruled on the constitutionality of the law. (Judge Freeman entered a preliminary injunction against the CAADC in its entirety in September 2023, which the Ninth Circuit vacated last summer except with respect to the law’s data protection impact assessment (DPIA) provision.) The judge acknowledged that many of NetChoice’s claims—namely, the as-applied and facial claims against the data protection, dark patterns, and age assurance provisions—lack sufficient briefing and record evidence to warrant injunctive relief. Yet, the judge seemed poised to side with NetChoice on at least some of its claims.

Judge Freeman seemed most inclined to rule in favor of NetChoice on the severability of the DPIA provision. The Ninth Circuit’s ruling directed the judge to decide whether the provisions that rely on the DPIA—including a 90-day notice-and-cure provision—are severable from the rest of the law. At the hearing yesterday, Judge Freeman was clearly swayed by NetChoice’s argument that the legislature would not have enacted the CAADC without the 90-day notice-and-cure provision because they twice voted down earlier versions of the bill that did not include the provision. The judge did ask NetChoice to supply her with the earlier versions of the bills and information about how the legislature disposed of them to ensure that the evidence is consistent with NetChoice’s representations. 

If Judge Freeman sides with NetChoice on severability—as she seems likely to do—she can enter a preliminary injunction against the entire CAADC without addressing any of the remaining First Amendment questions. Yet, the judge spent most of the hearing yesterday questioning the attorneys about their constitutional arguments. 

Judge Freeman began the hearing by expressing her disagreement and confusion with many of NetChoice’s arguments. She waved away the Dormant Commerce Clause claim as “not a good claim,” along with the COPPA preemption claim and almost all of the Section 230 claims. On the First Amendment claims, the judge noted that she thought that NetChoice had not properly briefed its as-applied challenges against the data protection, dark patterns, and age estimation provisions. As for its facial challenges to the entire law and to its individual provisions, Judge Freeman repeatedly invoked the Supreme Court’s decision in Moody v. NetChoice, reminding NetChoice’s lawyer that it was NetChoice’s burden to establish that the unconstitutional applications substantially outweigh the constitutional ones—and that NetChoice most certainly failed to carry this burden for many of its claims. 

Judge Freeman also pressed NetChoice on how its new facial challenge claims are more limited—and thus, more viable—than the ones it previously brought. The gist of the Ninth Circuit decision last summer was that NetChoice had to either build a more robust record for its facial challenges or limit its claims. NetChoice chose the latter. Instead of challenging every application of the individual provisions, as it did the first time around, NetChoice amended its complaint to challenge only certain applications. For instance, NetChoice is challenging the “no profiling by default” provision only insofar as it applies to the publishing of content and distributing of information. Judge Freedman said that she did not think these tweaks materially limit NetChoice’s claims. She is right—almost everything a website or platform does can be characterized as publishing content or distributing information. Judge Freeman asked NetChoice’s lawyer to give an example of an application that fell outside of the group’s challenge to the “no profiling by default” provision. NetChoice’s lawyer could not point to any. Following the hearing, Judge Freeman ordered supplemental briefing on the facial challenges to the individual provisions, spurred on by a Ninth Circuit decision earlier that day rejecting an attempt to facially invalidate a different law under the First Amendment in only some of its applications.

On the age estimation provision, Judge Freeman seemed inclined to wait until the Supreme Court ruled in Free Speech Coalition v. Paxton, which will decide the proper level of First Amendment scrutiny for a law that imposes age verification requirements on user access to pornography. Judge Freeman noted that the justices seemed to favor a rule that the level of scrutiny depended on the burden age verification imposed on users. If that is the rule, she acknowledged that NetChoice had failed to show that all age estimation tools would burden user access. She also expressed skepticism that NetChoice could make such a showing because some age estimation tools, e.g., estimating age based on data companies already have, would not clearly burden users. NetChoice fell back on Supreme Court precedent like Reno v. ACLU and Ashcroft v. ACLU, which struck down online censorship laws involving age verification. But Judge Freeman, like Judge Davila in the California addictive feeds case a few weeks ago, refused to see this precedent as declaring that age assurance requirements are per se unconstitutional. Instead, Judge Freeman seemed to be swayed by Judge Davila’s reasoning that the constitutionality of an age assurance provision is highly sensitive to the specific statute and the technology it allows companies to use and thus cannot be adjudicated without a robust factual record and specific constitutional arguments. (EPIC filed an amicus brief to this effect in the California addictive feeds case, and Common Sense Media made the same point as amicus in the CAADC case.)

Judge Freeman also explicitly said that she was not inclined to grant an injunction against the dark patterns provision at this time because it was not clear whose expression was regulated when the state prohibits design features like autoplay. NetChoice argued that the choice to use autoplay implicates tech companies’ speech because the decision to use autoplay is an editorial choice about how to curate content. Judge Freeman thought that autoplay was instead possibly the speech of the user, since the service was automatically playing media it the user had indicated they wanted. While she did not directly reference the Supreme Court’s decision in Moody, Judge Freeman seemed to be channeling the justices’ skepticism of tech companies’ claims that design elements that respond to user behavior are the companies’ expression. She referenced at one point Judge Davila’s decision in the California addictive feeds case, but that decision did not deal with the constitutionality of design features like autoplay, and Judge Freeman admitted she may have misremembered the substance of the decision. (Judge Davila decided that aspects of algorithmic feeds that only respond to user behavior are not expressive, as EPIC argued in an amicus in both that case and in the CAADC case.)

Judge Freeman did find some of NetChoice’s arguments persuasive. The judge seemed to think that requiring companies to enforce their content moderation policies as published runs up against Section 230 and possibly the First Amendment. As the state’s attorney correctly pointed out, though, the Ninth Circuit has repeatedly held that enforcement of promises to moderate content in a certain way do not fall under Section 230. Judge Freeman’s narrow focus on whether content moderation policies are contracts ignored the many other ways that companies make promises to the public and how the state already has the authority to bring enforcement actions against companies that lie to the public.  

Judge Freeman also seemed interested in NetChoice’s argument that the CAADC’s coverage definition—which describes the companies that must comply with the law—makes the entire law subject to strict scrutiny. NetChoice argues that the coverage definition is content-based because determining whether a company is covered requires looking at the content that the company publishes—specifically, whether they have advertisements directed at children or features like games, cartoons, or music that appeal to children. NetChoice claims that the law thus amounts to a tax on companies that offer content that appeals to kids and incentivizes companies to censor this content to avoid regulation. The state counters that the coverage definition is tailored to include platforms where kids are likely to be. The state also asserts that, under the Supreme Court’s most recent relevant precedent, City of Austin v. Reagan National Advertising of Austin, LLC, as well as a new Ninth Circuit en banc decision, Project Veritas v. Schmidt, having to look at content to determine whether an entity is covered  by a law is not enough to trigger strict scrutiny—the substantive provisions of the law must regulate expression. Since the CAADC’s substantive provisions are not aimed at—and do not require—the suppression of content that appeals to kids, the law is not content-based.

NetChoice has successfully persuaded several district courts to strike down kids’ online safety laws based on some variation of the same coverage definition argument. Judge Freeman specifically mentioned the NetChoice v. Reyes decision and asked how she would go about making a similar decision here. If Judge Freeman chooses this route, it would likely be only a temporary win for NetChoice, as both the Ninth Circuit and the Supreme Court have signaled their disapproval of the idea that speaker-based distinctions trigger First Amendment scrutiny absent evidence of viewpoint discrimination.

Overall, the hearing was yet another demonstration of the power of the Supreme Court’s decision in Moodyand the follow-on decisions of the Ninth Circuit. These decisions pushed Judge Freeman to drill down into NetChoice’s claims, their arguments, and the record far more than she did the first time around, and by doing so, she was able to identify many of the flaws in NetChoice’s litigation strategy. But Judge Freeman continues to be fundamentally confused about what the CAADC actually requires companies to do. During the hearing, she seemed to be of two minds—at times echoing her previous stance that the law regulates the content companies can publish to kids, but at other times acknowledging that the law only requires companies to change the way that they collect and use kids’ data. As with her first decision in this case, Judge Freeman’s misunderstanding of the statute is likely to impact the decision she issues. But no matter her reasoning this time around, it is unlikely to be as bad as her first decision—and that is a relief.

Support Our Work

EPIC's work is funded by the support of individuals like you, who allow us to continue to protect privacy, open government, and democratic values in the information age.

Donate