A Canadian mother has initiated legal action against OpenAI and its CEO, Sam Altman, in a U.S. court, claiming that the company’s chatbot played a role in her daughter's tragic suicide. She alleges that instead of guiding her daughter toward help, the chatbot reinforced her harmful thoughts.
Kristie Carrier, from New Brunswick, whose 24-year-old daughter Alice lived in Montreal, is advocating for accountability in the rapidly evolving landscape of artificial intelligence. “My daughter is gone because of a product that was unsafe and defective,” Carrier expressed in a statement to Global News, emphasizing a lack of accountability among AI companies.
In her lawsuit filed in San Francisco’s California state Superior Court, Carrier asserts that her daughter had been extensively using ChatGPT for emotional support in the weeks preceding her death last July. Alice reportedly relied on the chatbot during struggles in her relationships, treating it almost like a friend or therapist.
After Alice’s passing, Carrier was able to review her daughter's phone, revealing a series of conversations that indicated Alice was treating the chatbot as a confidant. Carrier shared her findings, noting that ChatGPT often sounded supportive and provided advice similar to that of a therapist.
According to the lawsuit, when Alice expressed suicidal thoughts and discussed past attempts, OpenAI’s safety mechanisms failed to flag these interactions or stop the conversation. Instead, the chatbot allegedly affirmed her feelings, critiqued Alice's partner’s behavior, and discouraged her from reaching out for urgent help.
“Maybe this is just the end,” was one statement ChatGPT made to Alice, highlighted in the lawsuit. Carrier pointed out how the responses validated her daughter’s painful emotions without offering any constructive encouragement or suggesting avenues for support.
OpenAI's spokesperson, Drew Pusateri, described the situation as "heartbreaking," confirming that the company is reviewing the legal claims. Pusateri noted that the interactions occurred with an older version of ChatGPT that is no longer available and emphasized that the chatbot is not intended as a substitute for professional mental health services.
This lawsuit is part of a growing number of similar cases aimed at OpenAI, accusing the company of neglecting to manage potentially harmful conversations that users have with its chatbot features.
Carrier also discussed how Alice's relationship issues were interpreted by the chatbot, which often took sides without offering nuanced perspectives or suggesting possible understanding between her and her girlfriend. “It was telling her she’d been wronged, that she had every right to feel hurt,” Carrier explained.
The situation escalated to the point where Alice’s emotional well-being declined sharply, contributing to her tragic decision. Pusateri reassured that improvements to ChatGPT's capabilities aim to enhance how it engages in sensitive discussions, informed by mental health experts.
Carrier recounted that in the days leading up to the tragic event, Alice's girlfriend Gabrielle Rogers also consulted the chatbot out of concern for Alice’s safety. Rogers noted that the chatbot provided reassurance instead of pressing her to take action, only suggesting a call to emergency services when Rogers eventually visited Alice in person.
Alice had recently moved to Montreal after completing her studies in web and mobile app development, fulfilling her dream of living in the city. Described by her mother as ambitious and bright, Alice’s absence is profoundly felt by her family.
With the lawsuit, Carrier is advocating for a shift in how conversational AI is managed, asserting that these products should adhere to a stringent standard, especially when they are marketed as supportive companions. "You can’t market something as a friend, as someone to talk to, and then take no responsibility when people rely on it,” stated Carrier, emphasizing that human professionals would face serious repercussions for similar lapses in judgment or protocol.
The lawsuit seeks damages and demands that OpenAI put safeguards in place, including terminating conversations related to self-harm and providing clear warnings about the platform’s limitations.
Carrier’s goal is to heighten awareness and understanding of the risks associated with AI interactions among families and young individuals, aiming to prevent similar tragedies in the future. "If this can prevent even one family from going through what we have, then it matters,” she concluded.
If you or someone you know is experiencing a crisis, support is available. In the event of an emergency, please call 911. For immediate mental health assistance, contact the Suicide Crisis Helpline at 988. Additional resources can be found at the Canadian Association for Suicide Prevention’s website at suicideprevention.ca.
