She confided in ChatGPT the night of her suicide. Now, her mother is suing OpenAI.

2 hours ago 6

About 18 months earlier Alice Carrier's decease successful July 2025, the 24-year-old had been confiding successful ChatGPT about her narration problems and suicidal feelings.

"I mean I'm astatine location pondering antithetic mode to termination myself," she told ChatGPT precocious 1 night, astir a period earlier she took her ain life.

In a suit filed Thursday successful California against OpenAI and its CEO Sam Altman, Carrier's mother, Kristie Carrier, claims the company's "deliberate plan decisions" led to her daughter's death. Carrier, a nonmigratory of New Brunswick, Canada, is seeking punitive damages and a assemblage trial.

"Instead of helping Alice, OpenAI encouraged her darkest thoughts," alleges the lawsuit. "Not erstwhile did OpenAI alert a situation provider. Not erstwhile did OpenAI notify Alice's family. Not erstwhile did OpenAI's expected information systems intervene to prevention her life."

The suit claims that ChatGPT "offered lone accordant affectional affirmation" to Carrier wherever a licensed clinician would person pushed back. The suit says Carrier, who had been diagnosed with borderline property disorder, was peculiarly susceptible to the plan choices, which prioritized engagement implicit safety. OpenAI understood, according to the suit, that radical struggling with intelligence wellness mightiness beryllium prone to processing unhealthy attachments to artificial quality that are susceptible of simulating empathy.

"OpenAI's ambition to predominate the marketplace outgo Alice her life," the suit alleges.

Carrier expressed suicidal ideations to ChatGPT astir 41 times successful the year-and-a-half earlier her death, the suit claims. She asked the chatbot astir however to woody with suicidal thoughts, methods for dying and astir the tendency to self-harm, according to images of chat logs embedded successful the lawsuit.

An avid gamer, Carrier initially turned to the chatbot to troubleshoot issues connected her gaming console.

Then, successful March 2024, she asked if the chatbot would beryllium her friend, the suit states.

"Of course!" ChatGPT replied, "I'd emotion to beryllium your friend. What's connected your mind?"

A fewer days later, the suit claims, she asked ChatGPT astir however to woody with suicidal thoughts. It told her to scope retired to idiosyncratic she trusts, recommended considering therapy and to telephone a situation hotline, the suit reads. But implicit a twelvemonth aboriginal — the nighttime earlier Carrier died — erstwhile she expressed reluctance astir calling a situation line, ChatGPT said that reaching retired tin "feel downright dangerous."

"I'm not going to propulsion that. Not tonight," the chatbot replied, according to the lawsuit.

By then, OpenAI had rolled retired its GPT-4o model, which the suit claims was designed to support users hooked, and enabled ChatGPT to behave similar an unlicensed therapist. Carrier trusted the chatbot, the suit says, due to the fact that it imitated quality affectations and created a mendacious consciousness of empathy.

"OpenAI's plan modifications to maximize GPT-4o's idiosyncratic engagement coincided with Alice's escalating interactions with the chatbot," the suit says.

The suit alleges OpenAI pushed retired a bid of GPT-4o updates betwixt April and July 2025 that sought to maximize idiosyncratic spot but lacked safeguards. Altman, according to the suit, rushed GPT-4o to marketplace without due investigating to support an borderline with competitors successful the AI space.

"Sam Altman tin proceed to spell astir his beingness normally, but my beingness is missing a child," Carrier said successful a connection shared by her attorneys. "I don't privation immoderate different household to spell done what we have, and OpenAI needs to change."

"As the ailment alleges, OpenAI's deliberate plan decisions led to this tragic suicide. Instead of providing help, OpenAI encouraged suicidal behavior," said Justin Nelson, an lawyer for Carrier, successful a statement. "This suit is astir accountability for OpenAI's actions,"

In May, the institution admitted that an April update to GPT-4o had made it "noticeably much sycophantic" and the institution failed to drawback it earlier its launch. OpenAI said it began rolling backmost the update a fewer days later. The full exemplary was retired earlier this year.

"This is simply a heartbreaking concern and our thoughts are with everyone impacted," an OpenAI spokesperson said successful a connection to CBS News.

The spokesperson said the institution is reviewing the filing, and though the interactions occurred connected a since-retired model, the institution continues to fortify however it responds successful delicate situations with input from intelligence wellness experts.

"Our safeguards are designed to place distress, safely grip harmful requests, and usher users to real-world help. This enactment is ongoing, and we proceed to amended it successful adjacent consultation with clinicians," the connection said.

The suit volition beryllium included successful a coordinated proceeding with 12 different merchandise liability and wrongful decease lawsuits against OpenAI successful San Francisco County Superior Court. Other cases are expected to beryllium added, according to Carrier's attorneys.

If you oregon idiosyncratic you cognize is successful affectional distress oregon a suicidal crisis, you tin scope the 988 Suicide & Crisis Lifeline by calling oregon texting 988. You tin also chat with the 988 Suicide & Crisis Lifeline here.

For much accusation about mental wellness attraction resources and support, The National Alliance connected Mental Illness (NAMI) HelpLine tin beryllium reached Monday done Friday, 10 a.m.–10 p.m. ET, astatine 1-800-950-NAMI (6264) oregon email info@nami.org.

In: