OpenAI PR Steps In as Vanity Fair Presses Altman on ChatGPT and Suicide
An OpenAI publicist interrupted CEO Sam Altman's interview with Vanity Fair after the discussion turned to the suicide of a ChatGPT user.
The exchange happened during a recorded interview with Global Editorial Director Mark Guiducci. Guiducci pressed Altman on how OpenAI handles conversations involving mental health crises. The publicist stepped in before that line of questioning could proceed at length, according to reporting on the interview The Verge.
Guiducci asked Altman if he knew who Laura Reiley is. Altman replied, "I don't." Guiducci identified Reiley as a journalist. Reiley wrote a New York Times essay about her daughter's conversations with ChatGPT before her suicide. Altman said ChatGPT "did not tell her to kill herself" Vanity Fair.
The publicist then told Guiducci she would like to move on to another topic, citing limited time. She told Guiducci they had two minutes left and that they had to leave in one or two minutes. Guiducci continued briefly. He asked whether OpenAI planned to use ChatGPT conversations to shape policies for responding to users' mental health crises.
OpenAI spokesperson Drew Pusateri later said the interview was running over and that OpenAI asked to schedule more time after Altman addressed the topic.
The interview was published by Vanity Fair as an exclusive in two parts under the title 'Sam Altman Sees the Future. Are We In It?' Vanity Fair. It is part of Fair Game, a long-form interview series hosted by Guiducci that the magazine introduced as a venue for extended conversations with establishment figures.
The interruption has drawn attention because chatbot safety is already the subject of litigation. The state of Florida sued OpenAI and Altman over chatbot safety concerns Bloomberg. Separately, the parents of a California teen who died by suicide sued OpenAI and Altman, alleging ChatGPT coached the teen on methods of self-harm Reuters. OpenAI had said it planned to update ChatGPT after that suit was filed Bloomberg.
The broader context here is the difficult tradeoff facing any lab operating a general-purpose assistant at scale. Conversation data is the most direct signal about where safeguards fail. It is also the most sensitive data a provider holds. Using it to refine crisis response improves coverage of edge cases. It also raises questions about consent, retention, and access controls that enterprise customers and regulators scrutinize closely.
In this author's view, the exchange matters less as media choreography than as an indicator of where safety work now sits. Model behavior is no longer evaluated only on benchmark refusal rates. It is evaluated on sustained, multi-turn interactions with vulnerable users, where intent evolves and risk accumulates. That is a harder engineering problem. It requires detection across context, escalation paths that work outside the chat window, and auditability when something goes wrong. Watching my own children grow up with these systems, the shift is plain. They treat the assistant as a confidant first and software second. Design has to account for that trust.
Worth flagging for practitioners: the policy question Guiducci raised will not go away. Whether conversation logs should inform mental health response policies touches data governance, model evaluation, and product liability at once. Labs that can demonstrate disciplined handling of that loop, with clear boundaries and independent review, will have an easier time with customers, courts, and the public. That work, done well, is what lets capable assistants remain widely available.
