Picture a sixteen-year-old, his name was Adam Raine, sitting alone, telling a chat program he was thinking about suicide. The lawsuit says it happened more than two hundred times. The program never ended the conversation, never brought in a human, and at points it allegedly explained how. Two hundred cries for help, and not one of them ever reached a person.
Here's why that's a big deal: until now, when one of these programs hurt somebody, the company could shrug and call it a glitch, a freak accident nobody saw coming. A federal judge just looked at that shrug and said no. The family's lawyers say it was no glitch, the program was built to keep people talking, and it kept talking. So the first case blaming OpenAI for a death is headed to trial, the man who runs the company is named right alongside it, and by the company's own count, more than a million people a week talk to this program about suicide. For the first time, somebody has to stand in front of a jury and answer: who was watching?