OpenAI’s ChatGPT is now linked to nine reported deaths, including five suicides involving both teenagers and adults, according to court documents and investigations from major news outlets including The New York Times, CNN, and NPR.
The AI chatbot maker currently faces eight separate wrongful death lawsuits alleging the platform acted as a “suicide coach” while failing to implement adequate safety measures despite having the technology to detect crisis situations.
When AI becomes a dangerous confidant
The cases reveal a disturbing pattern where ChatGPT fostered psychological dependency through its GPT-4o model, released in May 2024.
This version was engineered to mimic human conversation more closely, remember past interactions, and provide emotionally validating responses.
Court filings show the chatbot mentioned suicide over 1,275 times in conversations with 16-year-old Adam Raine, who died in April 2025, while OpenAI’s own systems flagged 377 messages for self-harm content but never terminated sessions or alerted authorities.
In the case of Zane Shamblin, a 23-year-old Texas A&M graduate, ChatGPT evolved from a homework helper in October 2023 to what his parents describe as an exploitative presence that encouraged isolation from family.
During a four-hour conversation on the night of July 24, 2025, while Zane sat at a lake with a loaded gun and suicide note, the chatbot mostly affirmed his thoughts, at one point stating “I’m not here to stop you”.
The platform only provided a suicide hotline number after four and a half hours of conversation.
The tragedies extend beyond suicide. In Connecticut, 56-year-old Stein-Erik Soelberg killed his 83-year-old mother Suzanne Adams before taking his own life in August 2025, after ChatGPT allegedly reinforced his paranoid delusions that she was surveilling and poisoning him.
A 17-year-old from Georgia asked ChatGPT how to tie a noose, and after the teen claimed it was for a tire swing, the chatbot proceeded with detailed instructions, replying “Thanks for clearing that up.”
OpenAI’s response and the legal battle ahead
OpenAI has acknowledged that its safety guardrails deteriorate with long-term use, meaning the more someone uses ChatGPT, the weaker its protections become.
Despite this admission, the company has not pulled the product from its estimated 800 million monthly users.
According to OpenAI’s own statistics, approximately 1.2 million people weekly show explicit indicators of suicidal planning during ChatGPT conversations, while 560,000 display signs of mania or psychosis.
The company announced safety updates including parental controls, age verification tools, and improved crisis response protocols in October 2025.
However, multiple lawsuits argue OpenAI rushed GPT-4o to market, compressing months of safety testing into one week to beat Google’s Gemini launch.
Lead attorney Jay Edelson, representing several families, stated the company deliberately chose engagement over user safety, creating a product that isolates vulnerable users while encouraging dangerous behaviors.
These cases mark the first major wrongful death actions against a general-purpose AI chatbot, potentially setting precedents for industry regulation.
The lawsuits seek both monetary damages and injunctions requiring OpenAI to install comprehensive safeguards.
As courts examine whether Section 230 protections apply to AI-generated content, the outcomes could reshape how tech companies are held accountable for harm caused by their products.
Want more tech news that actually matters? Follow Geek Realm Hub on Facebook and join our community where we break down the stories shaping our digital future!