What The Court Actually Decided
On August 10, 2026, the Ninth Circuit rejected an attempt by Meta, Google, TikTok and Snap to short-circuit thousands of lawsuits alleging their platforms were engineered to addict minors. The panel's holding was narrower than the headlines suggest: it did not rule that Section 230 of the Communications Decency Act fails to protect the platforms. It ruled that the companies asked the wrong court at the wrong time. Section 230, the panel held, operates as a defense to liability — something a defendant proves at trial or on summary judgment — not an immunity from suit that entitles a defendant to walk away before litigation even starts. Immunity from suit can be appealed immediately, mid-case; a liability defense normally can't. So the interlocutory appeal was premature, and the underlying design-defect claims go back to the district courts (Claims Journal, Aug. 10, 2026).
That sends roughly 2,400 federal cases — consolidated since October 2022 as MDL No. 3047, In re: Social Media Adolescent Addiction/Personal Injury Products Liability Litigation, before Judge Yvonne Gonzalez Rogers in the Northern District of California — back toward trial, alongside a parallel track of about 3,300 coordinated cases in California state court (JCCP 5255, Los Angeles County Superior Court, Judge Carolyn Kuhl) (Claims Journal, Aug. 10, 2026; Tech Policy Press JCCP tracker).
The Design-Defect Workaround
Section 230 says a platform can't be treated as the "publisher or speaker" of content posted by someone else — the provision that has shielded platforms from defamation, negligence and failure-to-warn suits over what users post (47 U.S.C. § 230). Plaintiffs' lawyers in MDL 3047 and JCCP 5255 have built their cases to route around that shield entirely: they don't argue that any specific post harmed a child. They argue the product — infinite scroll, autoplay, push notifications, recommendation engines tuned to maximize time-on-app — is a defectively designed good, no different in principle from a car with faulty brakes. As Judge Kuhl put it in the state proceeding, "the fact that a design feature like 'infinite scroll' impelled a user to continue to consume content that proved harmful does not mean that there can be no liability for harm arising from the design feature itself" (Tech Policy Press).
That theory already has a track record. A Los Angeles jury awarded $6 million in March 2026 against Meta and Google under a negligent-design theory in K.G.M. v. Meta/Google — the first verdict of its kind. And in New Mexico, a state court has hit Meta twice this year: a $375 million consumer-protection verdict in March, then a follow-on order from Judge Bryan Biedscheid on August 7 requiring Meta to pay $567 million into a teen mental-health treatment fund after finding the company created a "public nuisance" that causes harms that "migrate to the internet as a whole and, perhaps most concerning, to the real world" (Claims Journal, Aug. 7, 2026).
Steelmanning the Plaintiffs
The strongest version of the plaintiffs' case doesn't require believing Section 230 was wrongly decided. It requires believing that Congress in 1996 was legislating about hosting and moderating third-party content — a genuinely distinct problem from choosing which addictive engagement mechanics to bolt onto a product. Courts routinely hold manufacturers liable for how a product is built, independent of what a user does with it; a car company can't invoke free-speech-adjacent immunity because a driver's radio played offensive lyrics. If a platform's engineers deliberately tuned variable-reward notification schedules to maximize compulsive use among people whose prefrontal cortex isn't finished developing, treating that as a pure content-moderation question stretches Section 230 well past its statutory text. That's a coherent legal theory, not a pretextual one, and it deserves to be tested on the merits rather than dismissed at the courthouse door.
Why the Line Still Matters
The risk is that "design, not content" is a distinction with less of a difference than it sounds. A recommendation algorithm's entire function is selecting and arranging third-party content — that's what makes it engaging in the first place. Courts applying the Barnes test have long held that claims "inextricably intertwined" with content-publishing decisions belong under Section 230's shield precisely because letting plaintiffs relabel a publishing decision as a "design choice" would let every content-liability claim back in through the side door. If design-defect theories succeed here without a workable limiting principle, the next suits won't stop at engagement algorithms for minors — they'll reach ranking, search, and any recommendation system a plaintiff's lawyer can call a "feature" rather than a "publishing judgment."
What Comes Next
August 10 changes nothing about whether these claims ultimately survive Section 230 — it just guarantees they'll be tested through years of discovery and eventual trials rather than resolved on the pleadings. EFF has consistently argued that carving exceptions into Section 230 tends to entrench the largest platforms, since only they can afford sustained litigation at this scale (EFF, Feb. 2026) — a dynamic worth watching as smaller platforms face the same design-defect exposure without New Mexico-verdict-sized balance sheets. The eventual appellate rulings on the merits, not this procedural threshold decision, will determine whether "addictive by design" becomes a durable Section 230 exception or a theory that collapses once courts have to draw its outer edge.