Juries Are Doing What Congress Wouldn't: Holding Platforms Liable for How They're Built

NewsOnScale Staff

August 31, 2026

For three decades, Section 230 functioned as an almost impenetrable legal shield for social media companies, deflecting lawsuit after lawsuit by framing platforms as neutral conduits for user speech. That framework is cracking, not from legislative reform, but from courtrooms finding that the algorithms, recommendation engines, and engagement loops platforms deliberately engineer are products — and products can be defective.

Recent verdicts against Meta and Google represent something genuinely new in the legal landscape surrounding platform power. Plaintiffs — many of them families of children harmed by content rabbit holes, eating disorder promotion, or self-harm material — have successfully argued that the injury did not come from what users posted, but from how platforms were designed to maximize engagement with that content. That distinction matters enormously, and it may be the wedge that fundamentally reshapes what it means to operate a dominant digital platform.

## The Design Argument, Explained

Section 230 of the Communications Decency Act provides that platforms cannot be treated as the publisher or speaker of third-party content. Courts have historically interpreted this broadly, dismissing claims that would require a platform to have moderated differently. But plaintiffs' attorneys and legal scholars have spent years stress-testing a narrower question: what about the system the platform built to surface, amplify, and sequence that content?

A recommendation algorithm is not user-generated content. An autoplay feature is not a post. A notification system engineered to pull a teenager back to an app at 2 a.m. is a deliberate product choice. This is the terrain where juries have now found liability, and where Section 230's text — written in 1996, before algorithmic feeds existed — offers platforms far less cover than they have claimed.

The Massachusetts Supreme Judicial Court recently joined this legal current, ruling that design-based claims against social media companies are not automatically preempted by federal immunity law. Combined with the trial verdicts against Meta and Google, a pattern is emerging: the legal system is drawing a line between hosting speech and engineering behavior.

## What Platforms Built, and Why It Matters

The business model context is inseparable from the legal one. Engagement-maximizing design is not incidental to how Meta and Google generate revenue — it is the mechanism. Longer time on platform, more emotional arousal, more compulsive return visits: these are measurable outputs that platforms have optimized for, documented internally, and in some cases acknowledged cause harm to vulnerable users.

Internal research disclosed through litigation and regulatory inquiries has shown that engineers and policy teams at major platforms were aware of harms their design choices produced. That awareness, now part of the evidentiary record in multiple cases, is what transforms a legal theory into a compelling case before a jury. It is difficult to argue a design was reasonable when your own researchers flagged it as dangerous.

## The Reform Gap and the Litigation Substitute

Congress has held hearing after hearing on Section 230. The Senate Commerce Committee recently convened another, marking the law's 30th anniversary. Proposals to reform, narrow, or repeal the statute have emerged from both parties — often for contradictory reasons. Almost none have passed.

In the absence of legislative action, litigation has become the de facto policy instrument. This is inefficient, expensive, and accessible only to those with the resources to sustain multi-year legal battles against well-funded corporate defendants. Most harmed users never see a courtroom. The verdicts that do emerge are consequential signals, but they are not a substitute for durable regulatory clarity.

## What Comes Next

Platforms are already adjusting legal strategy, arguing strenuously on appeal that design claims are a backdoor to content liability and that the verdicts should be reversed. Some of those arguments may succeed. But the legal and cultural ground has shifted. Regulators, legislators, and — critically — juries now have a coherent framework for separating what platforms host from what platforms build.

For the AI agent economy emerging alongside these cases, the implications are immediate. If the systems an AI platform engineers to influence user behavior are products subject to product liability standards, the companies building autonomous agents that act on users' behalf are not insulated simply because an AI generated the output. The design question will follow these technologies wherever they go.