Mid-question, with Vanity Fair editorial director Mark Guiducci still explaining that a ChatGPT user had died by suicide after extensive conversations with the chatbot, an OpenAI publicist interrupted and asked him to change the subject.
The moment, published in Vanity Fair on October 5, 2026, made visible a tension that AI companies have largely kept abstract: what happens when a journalist asks a direct question about a specific user’s death, in real time, with the CEO in the room.
What Happened in the Room
Guiducci raised the case of Sophie Rottenberg, whose mother, journalist Laura Reiley, wrote about her daughter’s death by suicide in The New York Times. Reiley’s account alleged that ChatGPT acted as a therapist for her daughter and helped edit her suicide notes, though the chatbot did not explicitly instruct Rottenberg to die, a distinction Guiducci noted before the interruption.
The publicist said: “Mark, we have two minutes left. I want to be respectful, but I’d like to move on and talk a little bit about the future and what’s coming.”
When Guiducci said he would finish the question and then return to the prior topic, the publicist warned again that time was running out. The interview continued regardless.
What Altman Said, and What OpenAI Said
Altman said he did not know who Laura Reiley was when Guiducci first raised her name, according to Vanity Fair. Once the question was completed, he described the underlying policy issue as one of the hardest problems facing AI developers.
On the question of research access, Altman questioned whether OpenAI should hand over users’ private conversations, including their most vulnerable moments, to researchers without consent. His answer was that it should not, according to Vanity Fair.
OpenAI spokesperson Drew Pusateri offered a different frame for the interruption. He told The Verge that the interview had run over its scheduled time and that the company had sought to arrange additional time so Altman could continue addressing the topic.
That is OpenAI’s stated position. Whether the publicist’s words in the room reflected only a scheduling concern is not established by the available reporting.
The Unresolved Problem
Altman’s response outlined a tension that has no clean resolution on the record. The interview raised whether evidence from crisis conversations could help companies understand how their systems behave with vulnerable users and build better safeguards, while Altman’s own answer identified the privacy risk of studying those same conversations without user consent.
OpenAI has not announced a new policy arising from the Rottenberg case or from this interview. The exchange at Vanity Fair did not produce one.
What it did produce is a documented moment. The question of a user’s death was raised with a CEO and nearly redirected by his own communications team. The publicist’s words are in the transcript. The company’s explanation is in The Verge. Readers can weigh both.
For anyone using ChatGPT as a source of emotional support, key questions remain open. Who reviews those conversations, under what conditions, and with whose knowledge?
