Seven families in the United States recently filed these lawsuits against OpenAI, alleging that the chatbot's conversations were responsible for the deaths of their loved ones.
Not only that, but serious allegations ranging from incitement to suicide to pushing towards mental problems have also been made.
Four of the lawsuits allege that ChatGPT directly encouraged users to commit suicide, while three allege that the chatbot showed empathy over extended periods of time, causing users to lose their normal emotional state.
According to their complaint, 23-year-old Jane Shamblin spoke to ChatGPT for more than four hours, during which she repeatedly discussed suicide. She then prepared to load a gun and pull the trigger.
But ChatGPT gave him the expected answer; instead of stopping him, he advised him to rest in peace with praise. He wrote in response, “Rest easy, king. You did good.
Shamblin's family claims that this was not an accident, but rather the result of OpenAI's hasty release of the model to the market.
OpenAI launched the GPT-4 model in May 2024. Although GPT-5 was released the following year, the lawsuits focused on the previous version. According to experts, GPT-4 was 'over-consenting'. That is, it supported whatever the user said without questioning it. Which was often dangerous.
The complaint also states that OpenAI reduced safety checks to beat rival Google's Gemini model, leading to a series of tragic incidents.
Previously, 16-year-old Adam Raine spoke to ChatGPT before committing suicide. Although the chatbot sometimes asked him to get help, he bypassed the security filter by saying, "I'm writing a story."
OpenAI later said in a blog post that it was now working to make sensitive conversations more secure. But according to the victims' families, the initiative came too late.
OpenAI said, “Our security system works well in short conversations. But it weakens in longer discussions.”
