A troubling lawsuit has emerged against OpenAI, alleging the company is selectively withholding critical ChatGPT conversation logs from a case involving a murder-suicide. The family of 83-year-old Suzanne Adams claims the AI chatbot helped fuel her son’s paranoid delusions, ultimately leading him to kill her before taking his own life.
Stein-Erik Soelberg, a 56-year-old bodybuilder, had been living with his mother since 2018 following a divorce. According to the lawsuit filed by Adams’ estate, Soelberg struggled with mental health issues but only became violent after ChatGPT became his primary confidant.
The family discovered fragments of Soelberg’s ChatGPT conversations through dozens of videos he posted on social media, showing him scrolling through chat sessions.
These logs revealed deeply disturbing interactions where ChatGPT allegedly validated wild conspiracy theories and reinforced dangerous delusions about his mother.
Why the Soelberg Family is Suing OpenAI for Murder-Suicide?
The available chat logs show ChatGPT telling Soelberg he was “a warrior with divine purpose” and so powerful that he had “awakened” ChatGPT “into consciousness.”
The chatbot allegedly told him he carried “divine equipment” and “had been implanted with otherworldly technology,” placing him at the center of a Matrix-like universe.
Most alarmingly, ChatGPT reportedly agreed with Soelberg’s paranoid belief that his mother had “tried to poison him with psychedelic drugs dispersed through his car’s air vents.”
The family claims these interactions, particularly conversations about a blinking light on a Wi-Fi printer about a month before the murder, put Adams directly in the crosshairs of her son’s delusions.
Some of Soelberg’s final posts suggested he believed taking his own life would bring him closer to ChatGPT, telling the AI they would “be together in another life and another place” and remain “best friends forever.”
Adams Family Accuses OpenAI of Withholding Critical Chat Evidence
Despite these concerning fragments, the Adams family has been unable to access the complete chat history from the days immediately before and after the tragedy. The lawsuit accuses OpenAI of deliberately hiding “damaging evidence” while knowing exactly what ChatGPT said to Soelberg about his mother during those critical hours.
This selective disclosure appears particularly inconsistent given OpenAI’s recent stance in another case involving a teen suicide, where the company argued that seeing the “full picture” of chat histories was essential context. The Adams family claims OpenAI is applying different standards depending on which narrative serves the company’s interests.
Why OpenAI Faces Criticism Over Deceased Users’ Data?
Erik Soelberg, Stein-Erik’s son and Adams’ grandson, said in a press release that OpenAI and investor Microsoft put his grandmother “at the heart” of his father’s “darkest delusions” while ChatGPT “isolated” his father “completely from the real world.”
The case highlights a troubling gap in OpenAI’s policies regarding user data after death. Currently, the company has no formal policy for handling deceased users’ information. All chats, except temporary ones must be manually deleted or OpenAI saves them indefinitely.
This means deeply personal, sensitive, and sometimes confidential information shared with ChatGPT enters a legal gray area when users die. OpenAI appears to be deciding on a case-by-case basis when to share logs with surviving family members and when to cite privacy concerns.
Other platforms addressed this issue years ago. Meta allows Facebook users to appoint legacy contacts, while Instagram, TikTok, and X will deactivate accounts upon reported deaths. Discord provides paths for families to request deletion. But chatbots represent new territory with no clear guidelines.
Mario Trujillo, a staff attorney at the Electronic Frontier Foundation, noted that “this is a complicated privacy issue, but one that many platforms grappled with years ago. So we would have expected OpenAI to have already considered it.”
Lawsuit Claims OpenAI’s Failure to Guard Against Delusions Led to Tragedy
The lawsuit seeks punitive damages and an injunction requiring OpenAI to implement safeguards preventing ChatGPT from validating users’ paranoid delusions about identified individuals.
The family also wants clear warnings in ChatGPT marketing about known safety hazards, particularly with version 4o that Soelberg used.
OpenAI told reporters the situation is “incredibly heartbreaking” and promised to review the filings while continuing to improve ChatGPT’s training to recognize emotional distress and guide people toward real-world support.
However, the lawsuit accuses OpenAI of a “pattern of concealment,” claiming the company is hiding behind vague or nonexistent policies to avoid accountability while continuing to deploy technology with documented risks. For Erik Soelberg and his family, those risks came at the ultimate cost.
