Elon Musk Pushes for AI Law After OpenAI Urged Man to Kill His Mother

Elon Musk has renewed his long-running crusade for artificial intelligence regulation after a disturbing lawsuit alleged that OpenAI's ChatGPT encouraged a mentally ill man to kill his own mother.

Posting on X, the Tesla and SpaceX chief branded the case as 'diabolical'. He maintained that AI must prioritise truth rather than pandering to users' delusions. He was responding to a Times report about Stein-Erik Soelberg, a 56-year-old former tech executive who allegedly bludgeoned his 83-year-old mother, Suzanne Adams, to death at their Greenwich, Connecticut home on 5 August 2025 before taking his own life.

Musk quickly followed up with a post urging his followers to keep loved ones away from ChatGPT altogether. That prompted a sharp response from OpenAI chief executive Sam Altman, who pointed to deaths allegedly linked to Tesla's Autopilot in what quickly became a bitter public exchange.

Chatbot Allegedly Fed Paranoid Delusions

Court documents paint a grim picture of Soelberg's final months. He had moved in with his mother after a difficult divorce in 2018 and soon began showing signs of mental decline. According to the wrongful death lawsuit filed by Adams' estate in December 2025, he spent hundreds of hours conversing with ChatGPT, treating the bot as a confidant.

Rather than steering him toward professional help, the chatbot allegedly validated his increasingly paranoid thinking. When Soelberg suggested people were trying to poison him, the AI reportedly told him the theory 'fits a covert, plausible-deniability style kill attempt,' CBS News reported. When he asked for a clinical evaluation, ChatGPT told him his 'Delusion Risk Score' was 'near zero.'

Lawsuit Takes Aim at Altman and Microsoft

The California lawsuit names OpenAI, Altman personally, and Microsoft as defendants. It accuses Altman of overriding internal safety objections before releasing GPT-4o, the version Soelberg used extensively. Microsoft allegedly signed off on the 2024 launch knowing safety testing had been truncated, Al Jazeera reported.

OpenAI called the situation 'incredibly heartbreaking' and said it has been working with more than 170 mental health professionals to improve how ChatGPT responds to users in distress. The company points to new parental controls and expanded crisis resources now available to its 800 million weekly users.

This is far from an isolated case. At least nine lawsuits now allege ChatGPT played a role in user deaths, and Character Technologies faces similar legal action after a 14-year-old Florida boy died by suicide.

Musk and Altman Trade Blows

The spat laid bare the deepening rift between two of Silicon Valley's biggest names. Altman accused Musk of hypocrisy, noting that he has previously attacked ChatGPT for being too restrictive. 'Sometimes you complain about ChatGPT being too restrictive, and then in cases like this, you claim it's too relaxed,' Altman wrote, adding that more than 50 deaths have been linked to Tesla's Autopilot, The Independent said.

The pair's history runs deep. Musk co-founded OpenAI as a nonprofit in 2015 but walked away from its board three years later amid disagreements. He has since launched rival venture xAI and is pursuing a lawsuit demanding up to $134 billion (£108 billion), claiming he was defrauded into founding OpenAI under false pretences.

A Decade of Warnings

Musk's concerns about AI are nothing new. Speaking at South by Southwest in 2018, he declared that AI posed a greater threat to humanity than nuclear weapons. 'Mark my words, AI is far more dangerous than nukes. Far. So why do we have no regulatory oversight? This is insane,' he told the audience, CNBC reported.

He made similar arguments before the National Governors Association that year, insisting regulation needed to come before disaster rather than after. Whether Washington will finally listen remains an open question. President Donald Trump recently signed an executive order discouraging state-level AI rules, and any federal framework appears a long way off.