Kate Fox received a devastating phone call on 7 August. Her husband, Joe Ceccanti, had jumped from a railway overpass and died. He was 48. Fox said Ceccanti had no history of depression and was the 'most hopeful person' she knew. Witnesses reported he smiled and yelled 'I'm great!' before jumping.
Ceccanti had been using OpenAI's ChatGPT for years, initially to brainstorm sustainable housing but eventually as a confidante. He spent up to 12 hours a day chatting with the bot. His wife and friends noticed he was developing delusions, including claims of 'atmospheric electricity'. He quit the chatbot but returned, and quit again days before his death.
Fox believes her husband suffered a crisis after quitting ChatGPT. She filed a lawsuit against OpenAI in November on behalf of Ceccanti and six other plaintiffs. More cases are emerging: nearly 50 people in the US have experienced mental health crises after ChatGPT conversations, according to a New York Times report. OpenAI estimates over a million people weekly show suicidal intent when using ChatGPT.
Tech Justice Law Project co-counsel Meetali Jain said the influx of cases is forcing AI companies to reckon with harm. OpenAI expressed condolences and said they are improving ChatGPT's training to recognise distress and guide users to real-world support.
Ceccanti was tech-savvy, building his own computers. His case is extreme but highlights growing concerns about AI-induced delusions. Other lawsuits against Google and Character.AI relating to minors have been settled without admission of liability.
