A defense, not a wall: the 9th Circuit sends 3,000 youth-safety cases into discovery

On August 10 the 9th US Circuit Court of Appeals held that Section 230 of the Communications Decency Act "provides a defense to liability, not immunity from lawsuits, so the appeal was premature" ^[1]. That single procedural line keeps roughly 3,000 youth-safety cases against Meta, Google, ByteDance's TikTok and Snap moving forward, which means the companies keep producing documents ^[1][2].

The suits, brought by state attorneys general and by families, allege the four companies knew their products were addictive to children and teens and harmful to their mental health, and kept marketing them to young users anyway ^[2][3]. The companies had appealed a lower federal court ruling that let the cases be filed, arguing that Section 230 meant they could not be sued at all ^[4]. Section 230, now 30 years old, says platforms are not responsible for what their users post ^[5]. The distinction the 9th Circuit drew is the whole fight: a defense is something you win or lose on the record, after discovery, not a gate that closes before it.

The theory of these cases also sits awkwardly with the statute. They are less about what users posted than about how the platforms were allegedly engineered to present that content ^[6]. The Third Circuit has gone further than the 9th, ruling that Section 230 "does not provide immunity to platforms if they face tort lawsuits over injury caused by the algorithms they design" ^[7]. In other words, plaintiffs are treating the recommendation system as the product, and products can be defective.

The design argument has already produced money. A New Mexico judgment against Meta now totals $942 million after the judge added $567 million to the original award, finding that Meta had "created a public nuisance through its platform design" ^[8]. That puts the original figure at $375 million ^[9].

What discovery has surfaced so far explains why the companies wanted the gate to stay shut. According to reporting by Malwarebytes, a 2016 email attributed to Mark Zuckerberg said alerting parents to teens' live videos would "probably ruin the product from the start" ^[10]. A recent court filing alleged staff at social media companies compared their platforms to drugs, with one Meta employee writing "we're basically pushers" ^[11]. A filing in the New Mexico case says Snap employees were fielding roughly 10,000 sextortion reports per month by late 2022 ^[12]. An internal Snap investigation concluded 70 percent of victims never reported abuse because "they knew no action would be taken by Snap; indeed, of the 30% that did report, none were addressed" ^[13]. That figure came out through New Mexico's unredacted complaint, not from Snap ^[13].

The published safety features do not hold up much better under test. Researchers at NYU and Northeastern tested 86 youth-safety features and found 51 failed ^[14], a failure rate of about 59 percent ^[17]. By platform: Snapchat 73 percent, Instagram 66 percent, YouTube 55 percent, TikTok 50 percent ^[15]. Nine features could not be triggered at all, and every cyberbullying safeguard tested failed ^[16].

Watch whether the platforms seek Supreme Court review of the 9th Circuit's premature-appeal holding, since the circuit split with the Third Circuit on algorithm-design claims is the cleaner question for review ^[1][7]. Watch the New Mexico docket for further additions to the $942 million figure ^[8]. And watch what leaves the platforms' internal channels: every quarter these cases stay in discovery is another quarter of engagement research becoming public exhibits ^[11].