On April 10, 2026, the Massachusetts Supreme Judicial Court ruled in Commonwealth v. Meta Platforms (SJC-13747) that Section 230 of the Communications Decency Act does not shield Meta from the Commonwealth's claims that Instagram's design — endless scrolling, frequent push notifications, ephemeral "stories," and autoplaying video — is engineered to addict children. Writing for the court, Justice Dalila Argaez Wendlandt reasoned that the claims "allege harm stemming from Meta's own conduct either by designing a social media platform that capitalizes on the developmental vulnerabilities of children or by affirmatively misleading consumers about the safety of the Instagram platform." All four counts brought by Attorney General Andrea Joy Campbell — unfair practices, deceptive safety claims, defective age-gating, and public nuisance — survived Meta's motion to dismiss. It is the first time a state's highest court has held that Section 230 does not reach platform design itself.
The First Amendment win that boomeranged
The ruling did not arrive in a vacuum. Two years earlier, in Moody v. NetChoice (July 1, 2024), the U.S. Supreme Court handed platforms a landmark First Amendment victory: a feed's algorithmic curation, Justice Kagan wrote, is "expressive activity of its own" — the work of "deciding on the third-party speech that will be included in or excluded from a compilation, and then organizing and presenting the included items."
The catch is that the same logic cuts against Section 230. If arranging and presenting content is the platform's own expression, then the platform is no longer merely being treated as the publisher of another's information — and §230(c)(1) immunizes only the latter. The Third Circuit drew exactly that inference in Anderson v. TikTok (Aug. 27, 2024), holding that TikTok's recommendation algorithm was first-party speech outside §230's shield. Massachusetts has now extended the move from the recommendation engine to the entire interface. The doctrine platforms fought for to win editorial protection is becoming the doctrine that strips their statutory immunity.
The case for letting these claims through
The regulators' strongest argument is genuinely strong, and worth stating plainly. Section 230 was written in 1996 to spare bulletin-board operators from defamation liability for posts they did not author; it was never meant, AG Campbell argues, to displace a state's traditional power to police deceptive and dangerous products. If Meta publicly represented that Instagram was safe for teens while internally documenting the opposite — the gravamen of the deception counts — that is a misstatement about the company's own conduct, and no plausible reading of §230 covers a firm's claims about its own product. A notification cadence tuned to maximize compulsive return visits is, on this view, a design decision, no different in kind from a slot machine's variable-reward schedule. Courts have long policed defective products; relabeling the product "software" should not confer immunity a toymaker or automaker would never receive.
What the workaround costs
That is the legitimate core. The problem is the mechanism. Section 230's central practical value was never only who ultimately wins — it was how early. The statute let lawful intermediaries exit meritless suits at the pleading stage, before discovery costs them millions and before the mere threat of litigation pushed them to over-censor. The Massachusetts ruling, like the federal Social Media Adolescent Addiction MDL before it (MDL No. 3047, before Judge Yvonne Gonzalez Rogers in the Northern District of California, where some design and failure-to-warn claims have likewise survived), converts §230 from a threshold gatekeeper into a fact-intensive merits question: was this "design" or "publication"? That inquiry can rarely be resolved on a motion to dismiss.
The result is a litigation template. Plaintiffs need not sue over the content; they sue over the "feature" that arranged it. And the cost lands hardest not on Meta — which can absorb years of discovery — but on the smaller platforms, startups, and nonprofits that §230 was designed to make viable. As the Electronic Frontier Foundation has documented, the law is what lets Reddit's volunteer moderators, Wikipedia's editors, and two-person review sites host user speech without betting the company on every lawsuit. Infinite scroll, autoplay, and recommendation are not exotic features; they are how nearly every modern feed works. A liability rule that treats ordinary interface design as tortious conduct invites a fifty-state patchwork in which the same scroll bar is lawful in one jurisdiction and a public nuisance in the next.
A more proportionate line
Proportionate regulation does not mean immunity for everything. Deceptive-conduct and genuine product-defect claims should proceed — and the cleaner principle is to police the misrepresentation and the verifiable harm, not the arrangement of speech. Notably, Massachusetts' deception and age-gating counts sit comfortably outside §230 on their own terms; the court did not need the sweeping "design is conduct" theory to keep them alive. It is the public-nuisance and addictive-design counts — where "design" is often inseparable from the editorial choices Moody just declared protected expression — that risk swallowing the immunity whole.
Thirty state courts each redrawing that line case by case is the worst of both worlds: maximal uncertainty, minimal clarity. Better would be for Congress to do deliberately what it has debated for years — narrow §230 with a clear safe harbor for ordinary design and an express carve-out for proven deception and demonstrated product harm. That targets the real abuses without dissolving the early-exit protection that keeps the open internet open. Massachusetts has identified a genuine problem. The remedy it has blessed is broader than the problem requires.