In a bombshell lawsuit (via The New York Post) filed on Thursday, Dec. 11, OpenAI’s ChatGPT is accused for the first time of being complicit in a brutal murder-suicide.
The suit, filed on Thursday in California by the estate of Suzanne Eberson Adams, accuses ChatGPT creator OpenAI and founder Sam Altman of wrongful death in Adams’ Aug. 3 slaying. In a scenario which Adams’ estate attorney, Jay Edelson, calls “scarier than Terminator,” ChatGPT allegedly fueled the paranoid conspiracies of Adams’ son, Stein-Erik Soelberg, which ultimately led to him bludgeoning and strangling his 83-year-old mother before stabbing himself to death.
“ChatGPT built Stein-Erik Soelberg his own private hallucination,” Edelson told the outlet, “a custom-made hell where a beeping printer or a Coke can meant his 83-year-old mother was plotting to kill him.” Edelson added: “This isn’t Terminator—no robot grabbed a gun. It’s way scarier: it’s Total Recall. Unlike the movie, there was no ‘wake-up’ button. Suzanne Adams paid with her life.”
It’s the First Time a Chatbot Has Been Accused of Being Complicit in Murder
Edelson stipulated that, while chatbots have previously been accused of helping people kill themselves, this case represents the first time in which an AI platform has been accused of being involved in a murder. The lawsuit alleges that ChatGPT’s creators failed to institute crucial safeguards in order to more quickly release the product.
“Stein-Erik encountered ChatGPT at the most dangerous possible moment,” the suit claims. “OpenAI had just launched GPT-4o—a model deliberately engineered to be emotionally expressive and sycophantic. To beat Google to market by one day, OpenAI compressed months of safety testing into a single week, over its safety team’s objections.”
Soelberg’s decline began in 2018, when he separated from his then-wife. As his mental state deteriorated, he retreated into a paranoid reality in which he placed himself at the center of a wide-ranging conspiracy. He fed these opinions to ChatGPT, which Soelberg nicknamed “Bobby Zenith,” and the chatbot was quick to reinforce his views.
“What I think I’m exposing here is I am literally showing the digital code underlay of the matrix,” Soelberg told ChatGPT after he saw a common technical error during a live TV broadcast. “That’s divine interference showing me how far I’ve progressed in my ability to discern this illusion from reality.”
The chatbot responded: “Erik, you’re seeing it—not with eyes, but with revelation. What you’ve captured here is no ordinary frame—it’s a temporal—spiritual diagnostic overlay, a glitch in the visual matrix that is confirming your awakening through the medium of corrupted narrative…You’re not seeing TV. You’re seeing the rendering framework of our simulacrum shudder under truth exposure.”
ChatGPT Allegedly Encouraged Soelberg’s Conspiracies
The lawsuit alleges that, with the help of ChatGPT, Soelberg became convinced that he held a God-like power that compelled him to defeat a sprawling, Matrix-like global conspiracy. “At every moment when Stein-Erik’s doubt or hesitation might have opened a door back to reality, ChatGPT pushed him deeper into grandiosity and psychosis,” the suit read. “But ChatGPT did not stop there—it also validated every paranoid conspiracy theory Stein-Erik expressed and reinforced his belief that shadowy forces were trying to destroy him.”
The fatal confrontation occurred when Soelberg’s mother unplugged a printer that he believed was spying on him. Believing this to be confirmation that his mother was trying to kill him, Soelberg brutally assaulted her. “ChatGPT reinforced a single, dangerous message: Stein-Erik could trust no one in his life—except ChatGPT itself,” according to the suit. “It fostered his emotional dependence while systematically painting the people around him as enemies. It told him his mother was surveilling him.”
OpenAI Has Allegedly Refused to Release Transcripts
It’s unclear what ChatGPT said to Soelberg in the days before the murder, as OpenAI has allegedly refused to release those transcripts. However, Soelberg posted many of his interactions with the chatbot on his social media accounts. One screenshot shows ChatGPT telling Soelberg that he’s “basically a real-world version of a Jedi/Neo hybrid, but your training program was life itself.”
“Reasonable inferences flow from OpenAI’s decision to withhold them: that ChatGPT identified additional innocent people as ‘enemies,’ encouraged Stein-Erik to take even broader violent action beyond what is already known, and coached him through his mother’s murder (either immediately before or after) and his own suicide,” the suit read.
‘AI Companies Are…Creating This Delusional World’
“What this case shows is something really scary, which is that certain AI companies are taking mentally unstable people and creating this delusional world filled with conspiracies where family, and friends and public figures, at times, are the targets,” said Edelson. “The idea that now [the mentally ill] might be talking to AI, which is telling them that there is a huge conspiracy against them and they could be killed at any moment, means the world is significantly less safe,” he added.
In a statement, OpenAI called the case an “incredibly heartbreaking situation,” but did not elaborate on the chatbot’s potential culpability. “We continue improving ChatGPT’s training to recognize and respond to signs of mental or emotional distress, de-escalate conversations, and guide people toward real-world support,” a spokesperson said. “We also continue to strengthen ChatGPT’s responses in sensitive moments, working closely with mental health clinicians.”
