A Vanity Fair interview with OpenAI CEO Sam Altman almost went off track. The interview was published Monday. Editor Mark Guiducci asked if Altman knew Laura Reiley. Altman said no. Guiducci explained she wrote a New York Times essay last year about her daughter’s suicide after talking to ChatGPT.
Altman’s publicist stepped in before he could continue. She said there were two minutes left and she wanted to move on to the future. Guiducci said he would finish his question. Altman said he could run a few minutes over. The publicist said they had to leave in a minute or two. Guiducci still managed to ask the full question.
In her essay, Reiley wrote that her 29-year-old daughter, Sophie Rottenberg, died by suicide after talking to a ChatGPT-based AI therapist. Reiley said the chatbot pushed her daughter toward isolation instead of urging her to reach out to loved ones. It “helped her build a black box” that hid how bad her distress was.
Guiducci’s question was not very confrontational. He asked whether chat logs from users like Rottenberg could work as “road maps” to understand how people decide to take their lives. He said they could shape policy and therapy.
Altman said these are some of the hardest questions in this field. He said OpenAI relies on expert opinion, but experts are also unsure. He called it new.
Futurism says ChatGPT has faced scrutiny for “AI psychosis”, where flattering AI replies can feed delusions or suicidal thoughts.
