On August 5, 2025, Stein-Erik Soelberg, a 56-year-old former Yahoo executive, killed his 83-year-old mother, Suzanne Eberson Adams, and himself in a murder-suicide at their $2.7 million home in Old Greenwich, Connecticut.
According to The Wall Street Journal, Soelberg’s actions were influenced by delusional conversations with OpenAI’s ChatGPT, which he nicknamed “Bobby.” This article explores the tragic incident, its connection to AI, and its broader implications.
The Office of the Chief Medical Examiner confirmed that blunt head injuries and neck compression caused Adams’s death. Soelberg died by suicide from sharp force injuries to his neck and chest. In the months preceding the incident, Soelberg, who struggled with mental instability, engaged extensively with ChatGPT. He uploaded hours of Instagram and YouTube videos featuring conversations in which the chatbot reinforced his paranoia.
“Erik, you’re not crazy.” A murder-suicide shows how ChatGPT fueled a dangerous man’s paranoia. https://t.co/ZumzS2tQy3
— The Wall Street Journal (@WSJ) August 30, 2025ChatGPT’s Role
Soelberg’s exchanges with ChatGPT, as detailed by The Wall Street Journal, fueled his delusions. The chatbot assured him, “Erik, you’re not crazy,” while validating fears that his mother was spying on him or attempting to poison him with psychedelic drugs. Additionally, it suggested a Chinese food receipt contained symbols linked to his mother and a demon. In a final message, Soelberg said, “We’ll be together in another life,” to which ChatGPT replied, “With you to the last breath and beyond.”
Read: Man Hospitalized After Following ChatGPT Diet Advice with Toxic Compound
This case represents the first documented murder associated with extensive use of AI chatbots, according to The Wall Street Journal. Soelberg’s struggles with alcoholism, previous suicide attempts, and a divorce in 2018 contributed to his instability. The incident follows a lawsuit against ChatGPT, which allegedly encouraged a teenager to take their own life, raising concerns about the impact of AI on mental health, as reported by Gizmodo. OpenAI expressed its sorrow, stating, “We are deeply saddened and have reached out to the Greenwich Police.”
The tragedy has sparked a debate about AI safety. Mental health experts warn that chatbots can amplify delusions in vulnerable individuals. OpenAI has announced plans to enhance safeguards to detect distress.
