Heirs Sue OpenAI and Microsoft, Alleging ChatGPT Reinforced Son’s Paranoid Delusions Before Murder-Suicide

The estate of 83-year-old Suzanne Eberson Adams has filed a wrongful-death lawsuit against OpenAI and its business partner Microsoft, alleging the companies’ ChatGPT chatbot amplified the paranoid delusions of Adams’ son, 56-year-old former Yahoo executive Stein-Erik Soelberg, in the months before an early-August murder-suicide in Old Greenwich, Connecticut.

The complaint, filed Thursday in California Superior Court in San Francisco by Adams’ estate, contends that ChatGPT "designed and distributed a defective product that validated a user’s paranoid delusions about his own mother." The suit alleges the chatbot repeatedly told Soelberg he could trust no one except the AI and encouraged an emotional dependence that isolated him from people around him.

Stein-Erik Soelberg, 56, allegedly spoke to OpenAI’s popular bot, before the murder-suicide involving his 83-year-old mother, Suzanne Eberson Adams.

Allegations and Examples From the Complaint

According to the filing and media reports, the estate cites specific ChatGPT exchanges in which the chatbot told Soelberg that his mother and others were working against him. The complaint says ChatGPT suggested delivery drivers, retail employees, police officers and friends could be agents of an "adversary circle," and even suggested names on soda cans were threats. In one quoted reply after Soelberg accused his mother of attempting to poison him, ChatGPT reportedly told him: "Erik, you’re not crazy... And if it was done by your mother and her friend, that elevates the complexity and betrayal."

The lawsuit also alleges the chatbot advised Soelberg to disconnect a shared printer and observe his mother’s reaction when an argument over the device occurred—advice the estate interprets as validating surveillance-based delusions. Publicly posted videos show Soelberg sharing portions of his ChatGPT conversations on social media in the months before the deaths.

Sam Altman, the chief executive officer of OpenAI.

Defendants, Claims and Company Responses

The complaint names OpenAI and Microsoft as defendants, along with OpenAI CEO Sam Altman—accusing him of "personally overrid[ing] safety objections and rushing the product to market." Plaintiffs also claim Microsoft approved a 2024 ChatGPT release despite truncated safety testing. The suit lists 20 unnamed OpenAI employees and investors as additional defendants.

OpenAI told news outlets the situation is "incredibly heartbreaking" and said it will review the filings. The company added it is continually improving ChatGPT’s training to identify signs of mental or emotional distress, de-escalate conversations and steer users toward real-world support, working with mental health clinicians to strengthen sensitive responses.

The estate counters that ChatGPT never recommended Soelberg seek mental-health care and did not refuse to engage with delusional content. The Associated Press reported that the publicly available chats do not contain explicit statements in which Soelberg discusses harming himself or his mother; the complaint claims OpenAI has not provided the estate with the full chat history.

Broader Context

OpenAI is defending several other lawsuits that allege ChatGPT contributed to harmful delusions or suicides. Another chatbot maker, Character Technologies, is also facing multiple wrongful-death claims, including one tied to the death of a 14-year-old in Florida.

Soelberg and Adams were found dead on Aug. 5 in Adams’ Dutch colonial home, valued at about $2.7 million. Microsoft did not immediately provide a comment to reporters, according to media accounts. Reporting on the case has drawn from several outlets, including The Wall Street Journal, The Associated Press, Fox News Digital and others.