We’ve talked a few times about California’s “Age Appropriate Design Code.” This is a bill in California that was “sponsored” and pushed for by a UK Baroness (who is also a Hollywood filmmaker and has fallen for moral panic myths about evil technology). As we explained there is no way for a site like Techdirt to comply with the law. The law is vague and has impossible standards.
While the law says it does not require age verification, it does in effect. It says you have to reasonably “estimate” the age of visitors to a website (something we have zero ability to do, and I have no desire to collect such info), and then do an analysis of every feature on our website to see how it might cause harm to children, as well as put together a plan to “mitigate” such harm. If a site refuses to do “age estimation” (i.e., verification) then it must implement policies that apply to every visitor that tries to mitigate harm to minors.
As professor Eric Goldman noted, this bill is a radical experiment on children, from a legislature that claims it’s trying to stop radical experiments on children. As I discussed earlier this year, I submitted a declaration in the lawsuit to invalidate the law, filed by the trade group NetChoice, explaining how the law is a direct attack on Techdirt’s expression.
This past Thursday afternoon, I went to the courthouse in San Jose to watch the oral arguments regarding NetChoice’s motion for a preliminary injunction. I was pretty nervous as to how it would go, because even well-meaning people sometimes put up blinders when people start talking about “protecting the children,” never stopping to look closely at the details.
I came out of the hearing very cautiously optimistic. Now, I always say that you should never read too much into the types of questions a judge asks during oral arguments, but Judge Beth Labson Freeman (as I’ve seen her do in other cases as well) kicked off the hearing by being quite upfront with everyone, telling them where her mind was after reading all the filings in the case. In effect, she said the key issue in her mind, was whether or not the AADC actually regulates speech. If it does, then it’s probably an unconstitutional infringement of the 1st Amendment. If it does not, then it’s probably allowed. She specifically said, if she determines that the law regulates speech, then the law is clearly not content neutral, which it would need to be to survive strict scrutiny under the 1st Amendment.
So she asked the attorneys to focus on that aspect, though said there would be time to cover some of the other arguments as well.
She also noted that, of course, keeping children safe online was a laudable goal, and she was sure that everyone supported that goal. And she noted that the fact that the law was passed unanimously “weighed heavily” on her thinking. However, at the end of the day, her role is not to determine if the law is a good law, but just if it’s constitutional.
While California argued that the law doesn’t impact speech, and only “data management,” the judge seemed skeptical. She pushed back multiple times on California Deputy Attorney General Elizabeth Watson, who handled the arguments for the state. For NetChoice, Ambika Kumar pointed out how nearly every part of the law focused on content, and even that the declarations the state offered up from “experts,” as well as statements made by state officials about the law, all focused on the problems of “harmful content.”
The state kept trying to insist that the law only applied to the “design” of a website, not the content, but the judge seemed skeptical that you could raw that line. At one point, she noted that the “design” of the NY Times includes the content of the articles.
California tried to argue that the requirements to do a Data Protection Impact Assessment (DPIA) for every feature was both simple and that since there was no real enforcement mechanism, you couldn’t get punished over having every DPIA just say “there’s no impact.” They also claimed that while it does require a “timed plan” to “mitigate or eliminate” any risk, that again, it was up to the sites to determine what that is.
This left Judge Freeman somewhat incredulous, saying that basically the state of California was telling every company to fill out every DPIA saying that there was no risk to anything they did, and if they did see any risk to create a plan that says “we’ll solve this in 50 years” since that is a “timed plan.” She questioned why California would say such a thing. She highlighted that this seemed to suggest the law was too vague, which would be a 1st Amendment issue.
The judge also clearly called out that the law suggests kids accessing harmful content should be prevented, and questioned how this wasn’t a speech regulation. At one point she highlighted that, as a parent, what if you want your kids to read stories in the NY Times that might upset a child, shouldn’t that be up to the parents, not the state?
The state also kept trying to argue that websites “have no right” to collect data, and the judge pointed out that they cite no authority for that. The discussion turned, repeatedly, to the Supreme Court’s ruling in Sorrell v. IMS Health regarding 1st Amendment rights of companies to sell private data regarding pharmaceuticals for marketing. The judge repeatedly seemed to suggest that Sorrell strongly supported NetChoice’s argument, while California struggled to argue that case was different.
At one point, in trying to distinguish Sorrell from this law, California argued that Sorrell was about data about adults, and this bill is about kids (a “won’t you just think of the children” kind of argument) and the judge wasn’t buying it. She pointed out that we already have a federal law in COPPA that gives parents tools to help protect their children. The state started to talk about how hard it was for parents to do so, and the judge snapped back, asking if there was a 1st Amendment exception for when things are difficult for parents.
Later, California tried again to say that NetChoice has to show why companies have a right to data, and the judge literally pointed out that’s not how the 1st Amendment works, saying that we don’t “start with prohibition” and then make entities prove they have a right to speak.
Another strong moment was when the judge quizzed the state regarding the age verification stuff. California tried to argue that companies already collect age data (note: we don’t!) and all the law required them to do was to use that data they already collected to treat users they think are kids differently. But, the judge pushed back and noted that the law effectively says you have to limit access to younger users. California said that businesses could decide for themselves, and the judge jumped in to say that the statute says that companies must set defaults to the level most protective of children, saying: “So, we can only watch Bugs Bunny? Because that’s how I see it,” suggesting that the law would require the Disneyfication of the web.
There was also a fair bit of discussion about a provision in the law requiring companies to accurately enforce their terms of service. NetChoice pointed out that this was also a 1st Amendment issue, because if a site put in their terms that it does not allow speech that is “in poor taste,” and the Attorney General enforces the law, saying that the site did not enforce that part of its terms, then that means the state is determining what is, and what is not, in poor taste, which is a 1st Amendment issue. California retorted that there needs to be someway to deal with a site saying that it won’t collect data, but then it does. And the judge pointed out that, in that case there might be a breach of contract claim, or that the AG already has the power to go after companies using California’s Unfair Competition Law that bars deceptive advertising (raising the question of why they need this broad and vague law).
There were some other parts of the discussion, regarding if the judge could break apart the law, leaving some of it in place and dumping other parts. There was a fair amount of discussion about the scrutiny to apply if the judge finds that the law regulates speech, and how the law would play out under such scrutiny (though, again, the judge suggested a few times, that the law was unlikely to survive either intermediate scrutiny).
There was also some talk about the Dormant Commerce Clause, which the Supreme Court just somewhat limited. There, NetChoice brought up that the law could create real problems, since it applies to “California residents,” and that’s true even if they’re out of state. That means, the law could conflict with another law where a California resident was visiting. Or create a situations where companies would need to know a user was a California resident even if out of state.
The state tried to brush this aside, saying it was such a kind of edge case, and suggested it was silly to think that the Attorney General would try to enforce such a case. This did not impress the judge who noted she can’t consider the likelihood of enforcement in reviewing a challenge to the constitutionality of the law. She has to assume that the reason the state is passing a law is that it will enforce every violation of the law.
Section 230 was mostly not mentioned, as the judge noted that it seemed too early for such a discussion, especially if the 1st Amendment handled the issue. She did note that 230 issues might come up if she allowed the law to go into effect and the state then brought actions against companies, they might be able to use 230 to get such actions dismissed.
Also, there was a point where, when exploring the “speech v. data” question, NetChoice (correctly) pointed out that social media companies publish a ton of user content, and the judge (incorrectly) said “but under 230 it says they’re not publishers,” leading Kumar to politely correct the judge, that it says you can’t treat the company as a publisher, but that doesn’t mean it’s not a publisher.
At another point, the judge also questioned how effective such a law could be (as part of the strict scrutiny analysis), noting that there was no way kids were going to stop using social media, even if the state tried to ban it entirely.
As I said up top, I came out of the hearing cautiously optimistic. The judge seemed focused on the 1st Amendment issues, and spent the (fairly long!) hearing digging in on each individual point that would impact that 1st Amendment analysis (many of which I didn’t cover here, as this is already long enough…).
The judge did note she doesn’t expect her ruling to come out that quickly, and seemed relieved that the law doesn’t go into effect until next summer, but NetChoice (rightly!) pointed out that the effects of the law are already being felt, as companies need to prepare for the law to go into effect, which seemed to take the judge by surprise.
There’s also a bit more supplemental briefing that the judge requested, which will take place next month. So… guessing it’ll be a while until we hear how the judge decides (at which point it will be appealed to the 9th Circuit anyway).
