In this strangest of times, we are at a point in the genAI farce when the leading players are all competing with one another to explain just how dangerous their own tech is, and how close it is to wiping out humanity. As a play for further investment it’s certainly unprecedented, but it’s also proving deeply effective. In a new interview with Vanity Fair, Sam Altman is asked many questions about this peculiar time for his company, OpenAI, including one absolutely gut-wrenching moment in which Altman is confronted with the consequences of its incompetent LLM and his PR team desperately tries to close it down.
Anyone who actually uses LLMs knows firsthand that we are not on the verge of the singularity, where hyper-genius-level artificial intelligences start redesigning themselves at an exponential rate until humanity is nothing to them but a source of energy. They’re wildly broken guessing machines that regurgitate the internet back in beguiling and convincing tones, and the real and immediate danger is how many individuals become entranced by this often convincing parlor trick.
One such victim of this was the daughter of journalist Laura Reiley, a 29-year-old who tragically died by suicide after using Altman’s ChatGPT as a pseudo-therapist. While its programming, in this instance, stopped short of suggesting she kill herself, it had no qualms about agreeing to be her therapist, and went on to help edit her suicide notes. But when Vanity Fair reporter Mark Guiducci attempted to raise this situation with Altman, the CEO’s team interrupted in a desperate attempt to derail the conversation. Suddenly there were only two minutes left available, and that time, they decreed, was intended to be used to talk about the future of the company.
“So do you know who Laura Reiley is?” asked Guiducci.
“I don’t,” said Altman, before Guiducci went on to try to explain the story. Immediately a publicist piped up, “Mark? Two minutes left. I want to be respectful, but I’d like to move on.” And of course they would.
Mark Guiducci: “Do you know who Laura Reiley is?”
Sam Altman: “I don’t.”
Mark Guiducci: “She is a journalist. She wrote an essay in the ‘Times’ last year. Her daughter committed suicide after speaking to ChatGPT. ChatGPT did not tell her to kill herself—”
OpenAI Publicist:… pic.twitter.com/cE870oA3Se
— VANITY FAIR (@VanityFair) October 5, 2026
To Altman’s credit, he did not seem swayed by the publicist’s demands, and Guiducci remained determined, stating, “I’m going to finish this question.” Altman added that he could run a few minutes over. Not to Altman’s credit, he had no reasonable response to the situation. Guiducci wanted to know if, as Laura Reiley had suggested, the chat logs leading up to her daughter’s death could be used to better understand the problems with the software, and indeed to better understand therapeutic approaches to mental health.
“I mean, first of all, these are some of the hardest questions that we face or that anybody developing this technology faces,” replies Altman, in his usual way of appearing disarmingly honest while essentially shrugging his shoulders. “We try to rely on expert opinion as much as we can, but experts also aren’t sure what to do.” (Experts are sure. You’d better just program the LLM such that it refuses to participate in such conversations, or don’t release the LLM at all if it’s so obviously unsafe.) Altman continues, “The fact that someone in a moment of need that maybe can’t get help or isn’t willing to talk to a therapist can get some help here…Clearly people think there’s a version of this that is good. There are also clearly versions of it that are bad.”
It’s an astonishing statement in response to hearing, likely for the thousandth time, that his product helped edit a suicide note. There’s obviously no version of an unthinking, unknowing predictive text engine playing therapist to a real, living human that is good. There are versions that are lucky, certainly—I know someone whose kid got incredibly lucky when ChatGPT refused to help with suicidal ideation and instead responded that she could talk to her parents—but none that is good.
Altman then goes on to say that no, he does not think the chat logs (“private data”) should be used without consent (maintaining the laughable suggestion that they already aren’t), but adds, “Do we want to find ways where we can work with the community to address the mental health crisis in the world and for our products to be helpful there? Of course, yes.”
This section of the interview ends, bizarrely, with a game of word association. The first suggestion put to Altman: “Donald Trump.”
His reply?
“Powerful.”