Former xAI staffer says he was fired for questioning Grok safety

The xAI unit was at the centre of SpaceX's IPO roadshow, with Elon Musk touting AI satellites, data centres and a frontier lab that will help him create a vertically integrated AI conglomerate. Photo by LIONEL BONAVENTURE/AFP via Getty Images

Article content

Article content

In a lawsuit filed in California state court, Devin Kim claimed he repeatedly complained to his supervisors that the company was failing to prioritize safety while developing Grok. Kim says he was fired in September 2025, ahead of a planned presentation to xAI leadership about safety issues.

Article content

Article content

Article content

Kim’s claims come months after a global uproar over widespread usage of Grok to digitally undress people. Since then, regulators across Europe have sought to place limits on Grok. Some consumers, including influencer Ashley St. Clair, the mother of one of Musk’s children, have sued the xAI.

Article content

Article content

The company said in January that it disabled Grok’s ability to create sexualized images of real people and that it was “committed to making X a safe platform for everyone.”

Article content

In his suit, Kim said that, while working on a project, he found that Grok “tended to discriminate against certain racial groups while favouring others.” He said he flagged the observation to his supervisors in various reports. Kim said he was also concerned that Grok could give users guidance on bomb-making or creating bioweapons.

Article content

“Mr. Kim believed that xAI’s disregard for AI safety was unlawful,” his lawyers said in the lawsuit. Noting the undressing scandal, they added, “Mr. Kim’s concerns about xAI’s model were not only well-grounded but proved remarkably prescient.”

Article content

The complaint also names SpaceX, xAI’s parent company, which is set to have its blockbuster initial public offering on Friday. The AI unit was at the centre of the IPO’s roadshow, with Musk touting AI satellites, data centres and a frontier lab that will help him create a vertically integrated AI conglomerate.

Article content

Article content

Neither xAI nor SpaceX immediately responded to requests for comment on Kim’s suit.

Article content

Article content

Before joining xAI, Kim, who previously worked at Scale AI, claims that he spoke about AI safety with executives. During those conversations, Kim said he “highlighted the risks associated with rapidly advancing AI capabilities, and sought to understand xAI’s approach to safety and risk mitigation.”

Article content

But once he was at xAI, Kim claims his supervisor, co-founder Jimmy Ba, was dismissive of safety concerns. According to Kim, Ba didn’t follow safety directives issued by Musk.

Article content

“AI will kill us all anyway,” Ba allegedly said. Kim says it was Ba who fired him.

Article content

Ba, who didn’t immediately respond to an email seeking comment, exited the firm in February following its merger with SpaceX. Since then, all of xAI’s founding members except Musk have left along with dozens of engineers and specialists. Musk has since brought over staff from SpaceX’s Starlink unit.

Article content

—With assistance from Dana Hull.

Article content

Article content

Article content