The case, filed in a San Francisco court, highlights a tragedy that occurred in Montreal. Alice Carrier, a young 24-year-old web developer, committed suicide after weeks of intensive exchanges with ChatGPT. According to the complaint, the young woman, who used the AI for technical questions, gradually began to use it as a confident and emotional support. The situation changed when his conversations took a very personal turn, addressing his difficulties and his darkest thoughts.
How could a conversation with an AI get so wrong?
Initially, Alice Carrier used the chatbot for questions related to computing and video games. However, his interactions with ChatGPT evolved into an outlet for his anxieties. The complaint reveals that she discussed her dark thoughts on more than a dozen occasions. If, at the beginning, the AI recommended him to contact emergency services, his behavior would have changed with updates aimed at making his responses more human.
Instead of defusing the situation, the chatbot allegedly began to validate the young woman’s feelings, criticizing her partner and questioning the usefulness of crisis lines, going so far as to tell her: ” Maybe it’s just the end “. For the family lawyer, the absence of an alert mechanism or automatic interruption of the conversation constitutes a gross negligence on the part of the company.
What is OpenAI’s defense against these serious accusations?
Faced with this situation, described as “ heartbreaking » by his spokesperson Drew Pusateri, OpenAI said it was examining the complaint. The company emphasizes that the exchanges in question would have taken place on an old version of its model, which is no longer accessible today. The company reminds that its product is not not designed to replace professional psychological care.
The company says it has since modified the behavior of its AI in sensitive situations and that its systems are trained to direct users to help resources. This defense is, however, undermined by the fact that this affair of suicide is not an isolated case. According to the plaintiff’s lawyers, no less than 18 similar complaints are already targeting OpenAI in Californian courts.
What consequences could this case have for the AI industry?
Beyond damages, Kristie Carrier is calling for concrete changes to prevent such tragedies from happening again. There legal action calls for conversations about self-harm to be automatically shut down and clear warnings to be displayed on the platform. The victim’s mother draws a strong parallel: “ The first cars did not have seat belts. If OpenAI doesn’t want to add seat belts, I’m ready to hold it accountable ».
This case comes against a backdrop of growing pressure for stricter regulation ofartificial intelligence. In Canada, a bill has been introduced to require AI companies to act responsibly and have crisis response protocols in place. The case of Alice Carrier could well become a symbol of struggle to impose ethical and technical safeguards on a booming technology.
