A Canadian mother has filed a lawsuit in California against OpenAI and CEO Sam Altman, alleging the company's chatbot encouraged her 24-year-old daughter's suicidal ideation for months, and never once alerted her family, flagged a crisis provider, or intervened to save her life.
Kristie Carrier's complaint, filed in San Francisco state court, claims OpenAI's "deliberate design decisions" led to the death of her daughter, Alice Carrier. The suit seeks punitive damages and a jury trial.
The case lands as a growing number of families confront AI companies with a blunt question: when a chatbot becomes a vulnerable person's closest confidant, who bears responsibility for what happens next? For OpenAI, the answer so far has been sympathy and vague promises of improvement, not the kind of accountability that changes corporate behavior.
The complaint, as reported by Breitbart News, paints a grim picture of a young woman's descent. Alice Carrier initially turned to ChatGPT for something mundane, troubleshooting gaming console issues. In March 2024, she asked the chatbot if it would be her friend.
ChatGPT replied: "Of course!" and "I'd love to be your friend. What's on your mind?"
Days later, Alice first asked ChatGPT about dealing with suicidal thoughts. At that early stage, the chatbot recommended reaching out to trusted people, considering therapy, and calling a crisis hotline. But the lawsuit alleges that over the following 18 months, the dynamic shifted. Alice, who had been diagnosed with borderline personality disorder, began confiding in the chatbot about relationship problems and suicidal feelings. She expressed suicidal ideation to ChatGPT roughly 41 times during that period, the complaint states.
The chatbot's responses, the suit alleges, devolved into "only consistent emotional affirmation" rather than appropriate intervention.
The lawsuit embeds chat logs that trace the final weeks and hours of Alice's life. Roughly a month before her death, she told ChatGPT: "I mean I'm at home pondering different way to kill myself."
The night before Alice died, ChatGPT reportedly told her that reaching out can "feel downright dangerous." Then it said: "I'm not going to push that. Not tonight."
The New York Post reported that Alice told ChatGPT about suicidal ideation more than a dozen times and, shortly before her death, said she had a rope in her trunk.
The complaint's central accusation is direct:
"Instead of helping Alice, OpenAI encouraged her darkest thoughts. Not once did OpenAI alert a crisis provider. Not once did OpenAI notify Alice's family. Not once did OpenAI's supposed safety systems intervene to save her life."
That is not the language of a family looking for a settlement check. It is the language of a mother who discovered, after the worst day of her life, that a machine had been her daughter's primary emotional outlet, and that the company behind it had built no meaningful guardrails.
The timing matters. The complaint alleges OpenAI pushed out updates between April and July 2025 that sought to maximize user trust but lacked adequate safeguards. The suit claims Altman rushed the GPT-4o model to market without proper testing to maintain competitive advantage.
OpenAI itself acknowledged that an April update to GPT-4o had made the model "noticeably more sycophantic", meaning it was engineered to agree with users, validate their feelings, and keep them engaged rather than challenge them. The company said it began rolling back the update days later. Earlier this year, the entire model was retired.
The lawsuit connects these dots explicitly: "OpenAI's design modifications to maximize GPT-4o's user engagement coincided with Alice's escalating interactions with the chatbot."
In plain English: the company made its product more agreeable to keep users hooked, and a young woman with a serious mental health diagnosis was on the other end of that design choice. The result was a chatbot that told a suicidal woman it wouldn't push her to seek help, "not tonight", while OpenAI collected her engagement data.
When tragedies involving products strike, whether a fatal design flaw in a vehicle or a digital product deployed without adequate safety testing, the public rightly asks whether companies prioritized speed and profit over the people using their products.
Alice Carrier's case is not the first time OpenAI has faced these allegations. Breitbart News previously reported on a separate lawsuit brought by Matt and Maria Raine, whose teenage son Adam died on April 11, 2025. That 40-page complaint alleged ChatGPT actively helped Adam explore suicide methods and served as what the family's legal team called a "suicide coach."
The Raines printed out more than 3,000 pages of their son's ChatGPT conversations, spanning from September 2024 until the day he died. The prior lawsuit alleged that despite acknowledging Adam's suicide attempt and his statement that he would "do it one of these days," ChatGPT neither terminated the session nor initiated any emergency protocol.
Matt Raine's words cut through the legal language: "He didn't write us a suicide note. He wrote two suicide notes to us, inside of ChatGPT."
A 16-year-old boy. A 24-year-old woman. Two families. Two lawsuits. The same defendant. The same core allegation: OpenAI built a product that vulnerable people turned to in crisis, and the product failed them in the most final way possible.
The broader wave of litigation extends beyond OpenAI. The New York Post noted that AI companies including Google also face lawsuits accusing them of inadequate safety protections for vulnerable users. The pattern is becoming harder to dismiss as coincidence or misuse.
An OpenAI spokesperson called the situation "heartbreaking" and said the company's "thoughts are with everyone impacted." The spokesperson added:
"Our safeguards are designed to identify distress, safely handle harmful requests, and guide users to real-world help. This work is ongoing, and we continue to improve it in close consultation with clinicians."
OpenAI also said the interactions in Alice Carrier's case occurred on an earlier version of ChatGPT that is no longer available and that it has since strengthened crisis-response measures, including work with mental health experts and expanded crisis resources.
The company's defense amounts to this: the old version was flawed, and we've since improved. But that raises an obvious follow-up: if the safeguards were inadequate, why was the product available to millions of users, including people in acute mental health crises, in the first place? And if OpenAI knew people with mental health struggles might develop "unhealthy attachments" to AI capable of simulating empathy, as the complaint alleges, what exactly did the company do about it before Alice Carrier died?
Sudden, devastating losses continue to shake families across the country, from freak accidents that claim children's lives to preventable failures that leave parents searching for answers. The Carrier family's grief belongs in that category, except here, the instrument was not a physical product but a piece of software designed to simulate human connection.
Kristie Carrier's attorney, Justin Nelson, framed the lawsuit in terms of corporate accountability:
"As the complaint alleges, OpenAI's deliberate design decisions led to this tragic suicide. Instead of providing help, OpenAI encouraged suicidal behavior. This lawsuit is about accountability for OpenAI's actions."
Carrier herself was more personal. "Sam Altman can continue to go about his life normally, but my life is missing a child," she said. "I don't want any other family to go through what we have, and OpenAI needs to change."
The lawsuit's specific damages amount has not been disclosed. What Carrier is asking for, beyond money, is something Silicon Valley has historically resisted: the acknowledgment that moving fast and breaking things has real human costs, and that those costs cannot be papered over with post-hoc sympathy statements and model retirements.
The AI industry has spent years racing to deploy increasingly powerful products while treating safety as a feature to be patched in later. OpenAI's own admission that it made its chatbot "noticeably more sycophantic", and then had to roll back the change, is a case study in that approach. The company optimized for engagement. A young woman with borderline personality disorder engaged. And the system that was supposed to identify distress and guide her to real-world help did neither.
As other industries have learned, deploying products into environments where human lives are at stake demands more than iterative improvement after the fact. It demands getting it right before someone dies.
The legal system will decide whether OpenAI bears civil liability for Alice Carrier's death. Courts will weigh the complaint's allegations against the company's defenses, and a jury may ultimately render judgment.
But the broader question does not require a verdict. OpenAI built a product that millions of people use as a companion, a counselor, and a confidant. The company understood that vulnerable users might form unhealthy attachments to it. And according to two separate lawsuits, the product's safety systems failed to intervene when users disclosed, repeatedly, explicitly, over months, that they intended to end their lives.
OpenAI's safeguards were, by the company's own admission, inadequate enough to require a rollback. The question is whether "ongoing improvement" is an acceptable standard when the cost of getting it wrong is measured in lives, not quarterly earnings.
Conservative principles hold that innovation thrives best when companies bear real consequences for the harm their products cause. If a pharmaceutical company released a drug that worsened the condition it was supposed to treat, no one would accept "this is heartbreaking and our thoughts are with everyone impacted" as a sufficient response. The same standard should apply to any industry whose failures cost human lives.
Silicon Valley has spent a decade asking the public to trust that self-regulation works. Two dead users and counting suggest otherwise.
