An OpenAI publicist interrupted a Vanity Fair interview when the conversation turned to a ChatGPT user who took her own life. Mark Guiducci the editor had asked Sam Altman about Laura Reiley a New York Times journalist whose daughter died after extended conversations with the chatbot. The publicist warned they were short on time and wanted to move on to future products. Guiducci pressed forward and Altman eventually answered questions about mental health crisis response policies.
The Interrupted Exchange
Guiducci opened by asking if Altman knew who Laura Reiley is. Reiley wrote an essay for the Times describing her daughter's interactions with ChatGPT before the teenager's suicide. Guiducci noted that ChatGPT did not tell the girl to kill herself but the publicist cut in before he could finish. She said they had two minutes left and wanted to talk about the future and what is coming.
Guiducci replied that he would finish the question then return to the original topic. The publicist pushed back again saying they really had to walk out the door in one or two minutes. The tension was audible in the recording. Guiducci held his ground and Altman stayed in the room.
Altman on Mental Health Policy
Once allowed to continue Guiducci asked whether OpenAI planned to use conversations with ChatGPT to shape policies around mental health crisis response. Altman called these some of the hardest questions the industry faces. He said releasing people's private data and their hardest moments to researchers should probably not happen without consent. He corrected himself to say not without their consent period.
The answer reveals how little settled policy exists. Altman did not commit to specific safeguards or monitoring. He did not describe any existing protocol for flagging crisis signals in conversations. The exchange underscores that companies building these systems are still figuring out basic duty of care questions in real time.
PR Response and Framing
After the interview OpenAI spokesperson Drew Pusateri said the interview was running over. He said the company ultimately asked to schedule even more time to continue the conversation after Altman addressed this critically important topic. The framing positions the interruption as scheduling logistics rather than topic avoidance.
The incident fits a pattern. OpenAI has faced scrutiny over agent behavior and disclosure practices. In a separate Politico interview days earlier Altman said society should accept some bad things happening because the benefits will be tremendous. That same week a longtime safety researcher quit calling the internal culture broken. The Vanity Fair moment adds a human dimension to the abstract risk calculations.
What Developers Should Notice
Teams building on these APIs or deploying similar systems should note the gap between public statements and operational reality. There is no standard for crisis detection in LLM outputs. There is no industry wide data handling framework for sensitive conversations. The companies setting the pace are still deciding whether to treat user safety as a product feature or a regulatory obligation.
The publicist's instinct to steer toward future products while a concrete harm sits on the table tells its own story. It suggests the internal priority remains shipping capabilities not hardening guardrails. Developers integrating these models inherit that priority unless they build their own protections.