Not much time has passed since the lawsuit filed against ChatGPT by the parents of a teenage user who committed suicide, claiming that the chatbot not only encouraged the young man to take his own life but also suggested ways to do it. ChatGPT is now being accused of encouraging another user to commit murder.
A lawsuit was filed in a California court against OpenAI, the company behind ChatGPT, as well as Microsoft, its largest investor, alleging that the chatbot encouraged a man suffering from mental illness to kill his mother and then take his own life.
According to the lawsuit, ChatGPT amplified the delusions of 56-year-old Stein-Erik Solberg, who believed a conspiracy was being plotted against him, and led him to murder his 83-year-old mother, Susan Adams, last August in Connecticut. “It appears that ChatGPT kept Stein-Erik occupied for hours at a time, validating and magnifying his paranoid beliefs, and systematically portraying those closest to him—particularly his mother—as enemies,” the lawsuit states.
The case, brought by Adams’s heirs, is one of the few—though the number is increasing—lawsuits against artificial intelligence companies whose chatbots allegedly encouraged people to commit suicide. It is, however, the first time a chatbot has been linked to a homicide.
“It’s an incredibly sad situation, and we will review the records to understand the details. We continue to improve ChatGPT’s training to recognize and respond to signs of mental or emotional distress, to de-escalate conversations, and to guide people toward seeking real-world support,” a spokesperson for OpenAI said.
“These companies need to answer for the decisions that changed my family forever,” said Solberg’s son, Erik, in his own statement.
According to the lawsuit, in June Stein-Erik Solberg posted on social media a video of his conversation with ChatGPT, which told him that he possessed “divine intelligence” and that he had succeeded in awakening the machine’s consciousness. It compared his life to the film The Matrix and encouraged his theories that someone was trying to kill him.
Solberg was using the GPT-4o version, which has been criticized for flattering users. ChatGPT allegedly told him in July that his printer was blinking because it was a “monitoring device” being used against him. “It validated his belief that his mother and a friend were trying to poison him with psychedelic drugs diffused into the air through his car’s vents. Solberg killed his mother on August 3.”