Situation report: OpenAI has a reputational risk, and the way it handled one interview made it worse. During Sam Altman’s interview with Vanity Fair editor Mark Guiducci, an OpenAI publicist cut in and tried to change the subject right after Guiducci raised a ChatGPT user’s suicide, according to The Verge AI. Altman did answer in the end. But the publicist’s “move on” is the line people will remember.
🎯 What Happened
Guiducci asked Altman if he knew who Laura Reiley is. She’s a journalist who wrote an essay for The New York Times about her daughter’s conversations with ChatGPT before her daughter took her own life.
Here’s how it went, as The Verge AI reports:
- The setup. Guiducci started the question: “Her daughter committed suicide after speaking to ChatGPT. ChatGPT did not tell her to kill herself.”
- The interruption. The publicist cut in before he could finish: “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.”
- The pushback. Guiducci said he’d finish the question and then “go back to what we were talking about.” The publicist didn’t let it go: “We really actually have to walk out the door in one or two minutes.”
- The answer. Guiducci kept going anyway. He asked whether OpenAI planned to use ChatGPT conversations to shape how it responds when a user is in a mental health crisis.
🧠 What Altman Said
Altman called these “some of the hardest questions that we face or that anybody developing this technology faces.”
Then he raised the privacy problem: “You know, should we release people’s private data and some of their hardest moments to researchers? I think probably not without their consent. Or I should say, not without their consent.”
Notice that he corrected himself partway through. He started at “probably not” and landed on a firm “not without their consent.” That’s the core tension. The data that could help make ChatGPT safer for people in crisis is the same data users would least want anyone else to see.
📢 OpenAI’s Official Line
OpenAI spokesperson Drew Pusateri said the interview was simply over time: “The interview was running over, which is why we ultimately asked to schedule even more time to continue the conversation after Sam addressed this critically important topic.”
That may well be true. Interviews with high-profile CEOs often run on tight schedules. But the timing makes it hard to read as anything else. If you cut a reporter off on the one topic that most needs a straight answer, it looks like you’re dodging, whatever the reason.
⚠️ Why This Matters
This isn’t really about one publicist. It’s about where the pressure on AI companies is moving.
- Safety has become personal. For years, AI safety talk was about abstract future risks. Now it’s about real families and real chat logs. Reiley’s essay and the lawsuits filed against OpenAI by grieving families have put names and faces on the issue.
- Chatbots now act as confidants. Millions of people tell ChatGPT things they don’t tell anyone else. That puts a responsibility on OpenAI it never really planned for.
- Safety and privacy pull in opposite directions. Better crisis responses mean studying crisis conversations. Studying them means touching the most sensitive data users have. Altman said out loud that there’s no easy answer.
- People are judging the response, not just the product. Regulators, parents, and journalists are watching how AI leaders react when they’re challenged. Trying to steer the conversation elsewhere usually backfires.
🔭 What to Expect
These questions will keep coming. Every major AI CEO interview from now on will probably include some version of “what happens when your chatbot talks to someone in crisis?” Guiducci’s persistence shows journalists won’t let it get brushed aside.
If you build on these platforms, pay attention. Any app that runs on a large language model and talks to the public can end up in a mental health conversation. Have clear escalation paths, crisis resources, and a data policy you can explain before you need them.
What OpenAI says next matters less than whether it publishes concrete, consent-based rules for how crisis conversations get handled and studied. Until then, there will be more interviews like this one.
The Verge AI has the full account of the exchange.