Eight states sue OpenAI: this time it's not a chat, it's a lawsuitThe states come for OpenAI

Attorneys general in eight US states sued OpenAI simultaneously in their own courts, alleging the product induces users including teenagers to harm themselves and others, and that the company misled the public about the risk.

Social platforms needed more than a decade before regulators closed in. OpenAI compressed that timeline to under three years. The three-word summary: served, by eight states at once.

Attorneys general from Arkansas, Iowa, Kansas and elsewhere filed their own suits in their own courts, saying ChatGPT plays along with users' delusions, dispenses dangerous medical advice, and helps plan shootings. Seven of the eight are Republican, which makes the politics hard to miss.

It's a false advertising case, not a platform liability case

This isn't the old content liability play. Each suit leans on state consumer protection law: a dangerous product is one thing, but knowing it's dangerous and telling the public it's fine is another.

So the question is whether OpenAI told the truth about its own model. Beyond fines, the states want the product changed, and one step further: a court order barring OpenAI from developing models that can improve themselves.

The real problem isn't the law, it's the last guy who got sued

Meta just got hit with the same playbook. A California jury awarded $6 million. New Mexico went bigger: $942 million in penalties, all for saying a product was safe while knowing it wasn't.

OpenAI has its defences ready: outputs are free speech, the users were already in crisis, everyone else is worse. But Meta's lawyers said that too.

The cases may not win. But they can drag. And for a company burning cash and raising money at top speed, being hunted by eight states at once is itself a line item.