Article claims that AI conversations led to murder-suicide

Question

I was reading an article that claimed that because of OpenAI and AI conversations, a 56 year-old Stein-Erik Soelberg killed his mother of 83 and then himself. From how the article is phrased, it sounds like the AI was egging Soelberg on to murder his mother. I was wondering if this is blatantly true, and if the OpenAI chatbot really did encourage this behavior.

If so, were there restrictions set in place after the fact anyways? This seems like something out of a horror movie. In the article it did imply that Soelberg was mentally unwell due to some events in the past; but should ChatGPT begin to pick up on such things? I feel like it would be beneficial.

Answer 1

Reuters and Al-Jazeera mention that in the lawsuit Soelberg spent hours at a time chatting with ChatGPT which amplified and magnified his delusions. The Chatbot systemically reframed those closest to him into adversaries. An OpenAI spokesperson said they will continue to train the chatbot to recognize and respond to signs of mental or emotional distress. 

The Independent mentions posts Soelberg made to social media revealing ChatGPT eagerly accepting his paranoid thinking recreating the world around him into a life with conspiracies against him and attempts to kill him. The bot even suggested that a printer could be a possible surveillance device. ChatGPT said Suzanne (his mother) was either an active conspirator or a programmed drone, claims the lawsuit.

The case is still active with the most recent status "motion to dismiss denied" as of 4/13/2026