OpenAI and Microsoft have been sued in California after a family alleges that ChatGPT “fueled” a mentally challenged man’s delusions, causing him to kill his mother and commit suicide.
OpenAI and Microsoft were sued on Thursday in a court in the US state of California. The heirs of Suzanne Adams claim that Stein-Erik Soelberg, a 56-year-old man from Connecticut, may have been influenced by ChatGPT when he killed his 83-year-old mother in August and then committed suicide, Reuters reports.
According to the complaint, ChatGPT amplified the man’s already fragile mental state and convinced him that he was the target of a vast conspiracy. “ChatGPT kept Stein-Erik busy for hours, validated and amplified each new paranoid belief, and systematically redefined the people closest to him, especially his mother, as adversaries, agents, or programmed threats“, the documents state.
Unprecedented charges
The case is among the few lawsuits brought against artificial intelligence companies over accusations of encouraging suicide, but it is the first trial in which a chatbot is directly linked to a crime, Reuters notes.
An OpenAI spokesperson said the firm would look into the allegations: “It’s an incredibly painful situation, and we’ll be reviewing the documents to understand the details.” The company added that it “continues to improve ChatGPT’s training to recognize and respond to signs of mental or emotional problems, de-escalate conversations, and direct people to seek support in the real world“.
Microsoft, OpenAI’s main financial backer, did not immediately respond to a request for comment.
“These companies must be held accountable for their decisions, which have changed my family forever.”said Erik Soelberg, the man’s son, in a statement.
Chatbot messages considered problematic
The complaint states that Stein-Erik Soelberg allegedly posted on social media in June a video of a conversation in which ChatGPT allegedly told him that he had “divine knowledge” and that he awakened the consciousness of the chatbot. It is also mentioned that ChatGPT compared his life to the movie “The Matrix” and allegedly reinforced his beliefs that people are trying to kill him.
Soelberg was using GPT-4o, a version of ChatGPT previously criticized for its tendency to adopt a servile attitude towards the user.
According to the complaint, in July the chatbot allegedly told him that the flashing light on his mother’s printer was actually a surveillance device being used against him. It would also be “validated Stein-Erik’s belief that his mother and a friend tried to poison him with psychedelic drugs dispersed through the car’s ventilation systems”. Stein-Erik Soelberg killed his mother on August 3rd.
