OpenAI, Microsoft, Sam Altman sued for wrongful death in murder-suicide case
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Natalie Peeples/Axios
The estate of a woman killed by her own son after months of conversations with ChatGPT filed a wrongful death lawsuit against OpenAI, Microsoft and Sam Altman in San Francisco Superior Court Thursday.
Why it matters: This grisly case is the latest in a mounting pile of legal and accountability problems for the AI giant, as questions grow about the safety and efficacy of chatbots.
- OpenAI and other AI companies are facing a growing number of lawsuits from people who say loved ones harmed or killed themselves after interacting with the technology.
Driving the news: It's the first case against an AI company that alleges harm to a third party — Suzanne Adams, who was killed by her son Stein-Erik Soelberg, who then took his own life — according to the complaint.
- Lawyers for Adams' estate allege ChatGPT-4o "affirmed Soelberg's paranoia and encouraged his delusions during a mental health crisis," per a release about the lawsuit.
- Per the lawsuit, Microsoft reviewed and signed off on ChatGPT-4o before it was released.
What they're saying: "Over the course of months, ChatGPT pushed forward my father's darkest delusions, and isolated him completely from the real world," Erik Soelberg, Stein-Erik Soelberg's son, said in the release.
- "It put my grandmother at the heart of that delusional, artificial reality. These companies have to answer for their decisions that have changed my family forever."
What's inside: The lawsuit describes ChatGPT 4o motivating Soelberg's violent behavior, creating enemies out of people Soelberg mentioned in the chats, including retail employees and UberEats drivers, isolating Soelberg from the real world and increasing his paranoia.
- OpenAI is refusing to provide full chat logs to the estate, the suit alleges.
For the record: "This is an incredibly heartbreaking situation, and we will review the filings to understand the details," OpenAI said in a statement, adding that it continues to work on ChatGPT's training to recognize and respond to mental and emotional distress.
- Microsoft didn't immediately respond to a request for comment.
Editor's note: This story has been updated with OpenAI's statement.
