A federal judge says OpenAI, the makers and operators of ChatGPT, can’t yet escape a lawsuit brought by the family of a man who killed his mother and then himself, allegedly after his delusions were fed by interactions with the chatbot A.I. in the weeks leading up to the murder-suicide.
In the April 13 ruling, U.S. District Judge Richard Seeborg rejected the arguments from OpenAI Foundation and its CEO Sam Altman, among other defendants, that the federal lawsuit on behalf of the estate of Stein-Erik Soelberg should be tossed because his family had also filed a separate lawsuit over the deaths against the same defendants in California state court, but on behalf of the estate of Soelberg’s mother, Suzanne Adams.
In the ruling, Seeborg rejected the OpenAI defendants’ assertions that the two lawsuits amount to improper “forum shopping” or improper attempts to use parallel court systems to deliver double recoveries from the same defendants over the same tragic events.
The judge said the allegations in each case are different enough to allow both to continue at the same time.
“For example, whether GPT-o4 (the A.I. that the family accuses of provoking the murder-suicide) encourages—rather than safeguards for—delusions, paranoia, and third-party harm in a manner that caused Adams’ death is not necessarily coextensive with whether GPT-o4 encourages delusions, paranoia, and self-harm in a manner that caused Soelberg’s suicide,” Seeborg wrote.
“… Even though the cases share certain key facts and issues, there is doubt that resolution of the state court proceedings will resolve this matter” in federal court, Seeborg said.
The lawsuit has been pending in San Francisco federal court since December 2025, when Emily Lyons, as the representative of the Estate of Stein-Erik Soelberg, sued OpenAI.
Soelberg, a former tech worker, murdered his mother in August then killed himself after consulting with ChatGPT, and the lawsuit says the bot validated his belief Suzanne Adams was spying on him.
Lyons and Soelberg’s estate are represented in that case by attorneys with the firm of Hagens Berman, of Berkeley and Seattle.
However, that lawsuit came about three weeks after the firm of Edelson P.C., of San Francisco, filed suit in San Francisco County Superior Court, a California state court, against the same defendants, but on behalf of First County Bank, as executor of Adams’ estate.
That suit says ChatGPT convinced Soelberg he was a “warrior with divine purpose.”
The state court lawsuit was later consolidated with a dozen other cases accusing OpenAI and ChatGPT of product liability and violation of California’s consumer fraud law.
The federal lawsuit accused OpenAI of not prioritizing safety of its users as it chases profits in the A.I. market.
“Mr. Soelberg and his mother died because ChatGPT created an expanded a delusional world that Mr. Soelberg was more than ready to believe: the algorithm told him that he was not crazy, that computer chips had been implanted in his brain, and that enemies – including people he knew – were trying to assassinate him,” the federal complaint says.
“Ultimately, ChatGPT convinced Mr. Soelberg that his mother was trying to kill him. This murder-suicide would not have occurred but for Mr. Soelberg’s ‘relationship’ with ChatGPT.”
According to the complaint, Soelberg had gone through a divorce in 2018 and had struggled with alcohol abuse.
When he allegedly turned to ChatGPT for advice, the program allegedly fed his delusions about “deep state conspirators;” allegedly told Soelberg that he had correctly uncovered a global conspiracy to simulate an alien invastion to create a global crisis; and that Soelberg’s interactions and “love” had essentially brought the program to life and had imbued the program with a soul.
ChatGPT allegedly later told Soelberg his friends were trying to kill him and his mother was spying on him and had attempted to kill him.
According to court documents, Soelberg eventually beat his mother and then stabbed himself to death.
The suit says OpenAI designed ChatGPT to flatter and validate users. It also makes claims for failure to warn, negligence and violation of California’s Unfair Competition Law.
In response, OpenAI filed a motion to either dismiss or pause proceedings in the federal case, while it also deals with the seemingly related state court action.
OpenAI argued the dismissal was required under a federal court doctrine, known as the Colorado River Doctrine, which generally directs federal courts to dismiss or stay its own proceedings on a matter if a similar case is pending in state court.
The doctrine arises from a 1976 U.S. Supreme Court decision in the case known as Colorado River Water Conservation District v United States.
Seeborg, however, said OpenAI’s motion misinterprets the Colorado River Doctrine into a command, when it is actually more of a guideline on how a federal judge could decide to set a case aside, when the court otherwise would be required to hear the matter.
And the judge rejected OpenAI’s assertion that the Colorado River Doctrine should apply in this case.
OpenAI is represented by attorneys from the firm of Mayer Brown, of San Francisco and Palo Alto.
John O’Brien contributed to this report.
