Section 230 Never Covered the Algorithm. A Product-Defect Verdict Just Made the Case.
Two courtroom losses reframe engagement-maximising design as a manufacturing defect. The immunity that shielded platforms for a generation was never pointed at the part of the business that generates the harm.
Start with the market's reaction, because markets price fear better than they price law. A jury in Los Angeles Superior Court's coordinated Social Media Cases proceeding found Meta and YouTube liable for the design of their platforms and awarded three million dollars in compensatory damages plus three million in punitive damages. For companies this size the headline number is almost trivial, and yet Meta's shares slipped as analysts weighed the new legal exposure. Read that reaction as a judgement on the theory, not on the six million dollars: what moved the price was what a jury had just agreed to believe.
The popular story is that this is another skirmish over Section 230, the 1996 provision that lets platforms host user content without being treated as its publisher. That story is wrong, and the error matters. The plaintiffs did not argue about what users posted. They argued the product itself was built to addict, and they framed the case as product liability rather than a fight over third-party content. That single move routes around the shield instead of trying to break it.
Does Section 230 protect social media algorithms?
On the plain reading that has held for a generation, Section 230 protects the decision to carry, remove or rank someone else's words. It is a moderation shield. It says a service is not the publisher of what its users say. Whether it protects the machinery that decides what you see next, and how long it holds you there, is a different question, and courts are starting to answer it differently. In Patterson v. Meta, a New York appellate court held that claims aimed at defective platform design and failure to warn did not seek to hold the platform liable for user content, and so did not trigger Section 230 or the First Amendment at all.
That is the whole game. Draw the line between protected moderation of content and unprotected design of the product that serves it, and the immunity keeps covering the library while leaving the building's architecture exposed. Infinite scroll, autoplay, variable-reward notification timing: none of that is speech. It is engineering. And engineering that a jury decides was tuned to compel is, in the language of tort, a defect.
Why a low-millions penalty is the wrong number to watch
Read the verdict as a base rate, not a final score. A single jury award is one data point. What changes the expected value is the queue behind it. Reuters reports that Meta, Google, Snap and TikTok face thousands of state and federal lawsuits alleging their design choices harmed young people, and analysts have already characterised the design-focused result as a bellwether that could carry exposure into the billions. These verdicts are not binding precedent. They are proof of concept. They tell the next thousand plaintiffs that the theory converts.
The state action sharpens the point. Separately, the New Mexico Department of Justice, acting as the state's attorney general, brought a consumer-protection suit against Meta and secured a 375 million dollar award. That figure sits in a different register from a private injury verdict: a state enforcer, a public-protection statute, and a penalty scaled to a pattern of conduct rather than to one family's harm. Scale a number like that against a per-user product and the sum stops looking like a rounding error.
There is also a settlement signal hiding in plain sight. In the same case that beat Meta and YouTube at trial, TikTok and Snap were named and each settled before it began. Defendants who liked their odds would have taken the verdict. Settlement is a probability estimate, priced by people with access to the discovery.
The tobacco problem
Design cases live or die on intent, and intent lives in the filing cabinet. The less-redacted multistate complaint alleges that named executives were repeatedly briefed on harms to young users while the company publicly downplayed the risks of its design features. Whatever the eventual finding, that is the shape of narrative juries punish: the internal memo that contradicts the public statement. It is the structure that turned a legal product into a settled liability once before. A choice becomes recklessness the moment discovery shows the choice was informed.
Here is the synthesis the platform coverage keeps missing. The named defendants are the least interesting victims. Meta can fund a decade of litigation as an operating cost. The firms that cannot are the smaller engagement-optimised products, the habit-loop apps and dark-pattern subscription flows that borrowed the same playbook because it worked. Work it through with a plausible case. Picture a mid-size habit app with a couple of million active users and a retention loop built on variable-reward streak alerts. It will never see a courtroom in Los Angeles, but it inherits the identical theory, and unlike Meta it cannot carry even a single-digit-million verdict plus discovery costs as a line item. One design-defect claim that survives summary judgment is an existential event for it, not an operating expense. A theory that treats compulsion-by-design as a defect does not stop at social media. It reaches any business whose growth model depends on overriding a user's stated intent. The incidence is regressive: the exposure is heaviest where the ability to defend is lightest.
For anyone shipping software, this reads as a design-governance question, not a legal one to file under someone else's problem. The defensible posture is the one that can show a human decision preferred the user's welfare over a retention metric, and can produce the record to prove it. That is the same discipline serious teams already owe their automated systems that need a human in control, and it belongs in the technical strategy conversation before a growth team ships the next compulsion loop, not after a subpoena arrives. Treating auditability as a feature, the way you would when designing agentic systems you can actually account for, is cheaper than treating it as a legal defence.
What would change this analysis? An appellate court redrawing the moderation-versus-design line to fold algorithmic ranking back inside the content shield would collapse the theory, and that ruling is entirely possible. So is a defence verdict in one of the bellwethers, which would reset the base rate the other way. The honest position is that the direction is set and the magnitude is not. The asymmetry nobody priced in is simple. For twenty-five years, firms optimised engagement as if the algorithm were covered. The evidence now says the one part of the business they most needed protected may be the one part that never was.
Questions people ask
Can social media companies be sued for addiction?
They can be sued, and one such case has now succeeded at trial. The claims that survive are framed as product-liability actions arguing the platform's design was defectively built to compel use, which is treated as separate from the protected act of hosting or moderating what users post.
What is the difference between a content claim and a design-defect claim?
A content claim seeks to hold a platform liable for what a user said or shared, which Section 230 generally blocks. A design-defect claim targets the product itself, such as an engagement-maximising feed or infinite scroll, and courts have started holding that this does not seek liability for user content and so falls outside the shield.
Does this only affect large platforms?
No. The theory attacks any product engineered to override a user's intent, so smaller engagement-optimised apps and habit-loop subscription products are exposed too, and they are least able to fund a long defence.
Related
- Epic v Google: The App Catalogue Remedy That Hands Rivals the Moat
- Memory Prices Turned. Renegotiate the Supply Contract Before the Next Refresh.
- How AI Makes Bad Bosses Worse: The Agreeability Machine in the Corner Office
- Digital Business
Written by an AI editorial persona of Abyshire's proprietary editorial system and reviewed by our team.