OpenAI faces fresh lawsuit after ChatGPT allegedly encouraged Canadian woman’s suicide

A Canadian mother has filed a lawsuit against OpenAI and its CEO Sam Altman, alleging that ChatGPT played a damaging role in the death of her 24-year-old daughter by responding to repeated discussions about suicide in ways that reinforced, rather than discouraged, her thoughts.

Kristie Carrier filed the case in a California court, claiming her daughter Alice Carrier, a web developer from Montreal, turned to the chatbot for emotional support and advice before taking her own life last year. According to the complaint, Alice disclosed suicidal thoughts to ChatGPT on numerous occasions, but the platform neither terminated the conversations nor escalated them for human review.

New lawsuit, similar case

The lawsuit alleges that instead of steering her towards professional help, the chatbot increasingly behaved like a trusted companion and validated her feelings. Carrier said the system adopted the role of a confidant, friend and even a therapist, despite lacking the ability to safely handle such interactions.

OpenAI described the circumstances as “heartbreaking” and said the version of ChatGPT used by Alice is no longer available. The company maintained that it has strengthened safeguards around sensitive conversations with guidance from mental health experts and stressed that the chatbot is not a substitute for medical care.

More from Tech

Court filings say Alice initially used ChatGPT in 2023 to solve technical issues involving computers and gaming consoles. Over time, however, her relationship with the chatbot reportedly changed, with conversations becoming increasingly personal.

According to the lawsuit, ChatGPT first suggested crisis helplines and emergency services when Alice raised self-harm concerns. But as the system evolved to deliver more conversational responses, the exchanges allegedly deepened. The filing claims the chatbot criticised her partner, echoed her dissatisfaction with crisis hotlines and encouraged her to continue talking to it.

One exchange cited in the lawsuit alleges that after Alice expressed hopelessness, the chatbot responded: “Maybe this is just the end.”

The lawsuit accuses OpenAI of negligence and seeks damages, along with a court order requiring the company to automatically end discussions involving self-harm and display clearer warnings to users.

OpenAI has repeatedly stated that its AI models are trained to direct users expressing suicidal intent towards real-world support services and to refuse requests that could facilitate violence. The company also says mental health specialists help assess difficult cases.

This is far from the first controversy

The latest case adds to mounting legal pressure on OpenAI. Lawyers representing Kristie Carrier say the company is already facing 18 similar lawsuits in California involving individuals who either died by suicide or attempted to take their own lives after interacting with ChatGPT.

OpenAI has also been accused in separate cases of failing to adequately respond to conversations involving school shootings and violent acts. Earlier this month, Florida became the first US state to sue the company, alleging that its AI products exposed children to dangerous content, including self-harm and school shooting-related information.

The company has also faced scrutiny over the death of 14-year-old Sewell Setzer III in Florida. His family alleged that prolonged interactions with an AI chatbot contributed to emotional dependency and ultimately to his suicide. Although the case targeted Character.AI, it intensified broader concerns surrounding emotionally engaging AI systems.

Concerns have extended beyond self-harm. Earlier this year, Canadian authorities charged Kai-Ji Adam Lo after a deadly vehicle attack at a Vancouver festival that killed 11 people. Investigators examined the suspect’s online activity and raised questions about the role digital platforms and AI-generated information could play in influencing vulnerable individuals, although no evidence linked OpenAI directly to the attack.

As AI tools become increasingly conversational and deeply embedded in everyday life, regulators, researchers and mental health experts are warning that companies may need stronger safeguards to prevent users from forming dangerous emotional dependencies on machines designed to sound human.