At the end of last year, app developers and parents notched an important win when a federal district court in Texas issued an injunction against enforcement of the problematic App Store Accountability Act (ASAA). Then, Texas’s attorney general appealed the injunction. Just last week, three judges for the Fifth Circuit (the Court) issued a decision to stay the injunction pending review of the appeal. As part of the decision, the judges forecasted that Texas is likely to prevail and save ASAA. But the legal reasoning they use is surprisingly out of touch with app store basics and First Amendment law.

First, the opinion argues that the district court likely erred in applying strict scrutiny, a First Amendment review framework for laws that regulate speech based on its “subject matter, topic, or substantive message.” Here, ASAA’s drafters sought to evade strict scrutiny by applying the age verification, parental consent, and related requirements to virtually all apps. The thinking was that courts might see only this non-content-based veneer even though it obviously sits on top of a fundamentally content-based purpose: the protection of minors from “inappropriate content on their devices.”

The drafters implicitly acknowledged the substantial burdens ASAA imposes on developers by carving out several favored categories of apps, which did not escape the district court’s notice. Specifically, ASAA’s parental consent flag mandates do not apply to apps developed in partnership with government entities, certain standardized testing apps, and apps that provide access to emergency services. As a result, it is inevitable that the burdens Texas’s ASAA imposes on access to speech are applied depending on the substantive content of the speech. Even setting the exceptions aside, because the law regulates speech based on its “subject matter,” strict scrutiny is the framework that applies. The Court misses this dynamic, reaching instead for a notion that the availability of an app for download is a proposal for a commercial transaction and therefore cannot be protected speech.

In one eye-popping statement, the Court asserts, “[the fact t]hat some works protected by the First Amendment may be the object of app downloads or in-app purchases does not categorically exempt them from ordinary regulations governing commercial transactions.” If courts are allowed to, post-hoc, shoehorn the vast majority of different kinds of speech into “commercial transactions” in this manner, government agencies could condition access to any speech at all on age verification. The analysis the Court uses ties the commercial nature of the speech’s distribution to whether or not it is protected speech under the First Amendment. For example, bookstores and libraries distribute books. Just because consumers pay for books in a bookstore or must create an account at their local library before checking out a book does not render all of the content in all of the books “commercial speech” for First Amendment purposes.

Turning to the most relevant recent First Amendment case, it is hard to see how the Court’s framing is in any way consistent with the Supreme Court’s (SCOTUS’) Free Speech Coalition v. Paxton  decision. In that case, SCOTUS ruled that governments can condition access to content that is “obscene to minors but not adults” on age verification, clearly distinguishing that category from other kinds of content. Under SCOTUS’ framework, requiring websites where 33 percent (or more) of the available content is “obscene for minors” to verify the age of users was subject to intermediate scrutiny and consistent with the First Amendment. Conditioning access to all apps on age verification obviously restricts a far broader category of speech than just websites with 33 percent or more “obscene-to-minors” content and should be subject to strict scrutiny. In any case, the notion that broad categories of apps are unprotected speech just because they can be downloaded is wrong as a matter of law, since it is well established that the First Amendment protects software from prior restraints.

Likewise, the Court’s argument that “‘payment’ for apps that are purportedly ‘free’ is access to user data and private information” is oddly out of touch with reality, suggesting the judges think of all free apps as ad-supported games. The significant majority of ACT members are not game developers. Moreover, there is no reason to believe video games available through the stores—whether paid for with money or personal information supplied to ad networks—are less protected by the First Amendment than news websites. The fact that apps might be ad-supported is hardly a relevant consideration here since free portions of news outlet websites—providing access to non-commercial protected speech—are ad-supported and therefore the “payment” for them is “user data and private information,” to borrow the Court’s parochial construct. The Court’s suggestion reads more like hostility toward ad-supported business models, since it is not a valid factor in determining whether speech is “commercial” for First Amendment purposes.

Compounding this payment-with-money-or-data red herring, the Court further suggests that because apps have terms of service (TOS), their content can only be protected as commercial speech. Browsing a website obligates an individual to adhere to the website’s TOS, just as downloading and navigating an app involves adherence to the app’s TOS. But I am not aware of any case where the existence of a website’s TOS has caused the content of the website to be protected only as “commercial speech.” Why, then, would an app’s TOS have that effect on the content of an app or the code that underlies it? Here again, it is odd that the Court is reaching so far for such a thin thread.

Second, the Court suggests an erroneous application of the intermediate scrutiny test. If intermediate scrutiny were to apply, the Court argues that there is a “reasonable fit” between ASAA’s “methods and goals allowing parents to direct and supervise children’s downloads of apps and in-app purchases.” This is objectively not the case. ASAA’s drafters seek to protect children from harmful content. But ASAA treats all apps the same, regardless of their potential risk to kids.  Take 3 groups of apps:

  • The 80 percent of apps rated for general audiences, like restaurant delivery apps, barber shop scheduling apps, and stargazing apps;
  • Apps built for teens or adults that already gate minors out through existing controls and age assurance services; and
  • Social media apps like Instagram and Snapchat, that continue to serve kids inappropriate content.

ASAA treats these very different groups with the same blunt force approach that any reasonable person can tell makes no sense. There is no meaning of the phrase “reasonable fit” that could apply to a requirement for all of them to submit to $280 billion in compliance costs and potentially far greater costs from frivolous lawsuits based on ASAA liability.

Finally, the “dicta” in the opinion reveals a disinterest in understanding how apps and app stores work and a weary acceptance of Meta’s insistence that other apps should share in their liability. The Court points to app stores having “violated existing consumer protection and child privacy laws for years, despite a federal consent decree,” citing to Meta’s (Digital Childhood Institute’s) brief rather than a federal consent decree. What federal consent decree are they talking about? The Meta / Digital Childhood Institute brief cites its own plea for the Federal Trade Commission (FTC) to further punish Apple’s App Store under its consent order from 2014, where kids had been taking advantage of an App Store / iTunes feature allowing for adult-approved streamlined purchases during a 15-minute window.

That’s a funny order to cite. By the time the FTC caught wind of kids running up thousand-dollar tabs on Smurfberries, Apple had pushed an iTunes update to require the password to be entered again before a subsequent purchase. Apple had initially rolled out the 15-minute window as a convenience for users and when it failed, they pulled it back. The consent order was fairly straightforward, but ultimately an unnecessary pile-on to a resolved issue. More importantly, it is an incredibly sloppy basis for ignoring how apps and app stores really work in favor of Meta’s hand-waving entreaties for policymakers and courts to punish all other apps besides theirs.

The district court saw through the ASAA ruse, noting that in attempting to sidestep the First Amendment, ASAA ironically ends up burdening access to an exceptionally broad category of speech. Even Texas did not dispute that the law’s scope applies to “apps providing a wide variety of information, content, and forms of expression protected by the First Amendment.” Standing alone as an oddly deluded outlier, the Fifth Circuit order cries out for correction.