Former xAI engineer sues Musk’s company over Grok safety concerns

A former xAI engineer says Elon Musk’s company fired him after he raised alarms about Grok’s safety risks, turning a fight over AI guardrails into a California courtroom battle. Devin Kim’s lawsuit says xAI ignored repeated warnings that its chatbot approach could foster discrimination and even contribute to the proliferation of weapons of mass destruction.

Kim filed the complaint in California state court on June 10, naming both xAI and SpaceX. The suit says he pressed the company on the risks of moving too fast on Grok and was punished for doing so, a claim that places employee dissent at the center of the case. xAI and SpaceX did not immediately respond to requests for comment.

The filing also underscores how quickly Kim has moved from inside Musk’s orbit to the policy world around it. The Center for AI Safety announced on June 2 that Kim would become its president and that it was establishing the Frontier Security Institute in Washington, D.C. That puts him at the helm of an organization focused on AI safety just as his allegations against xAI ask whether companies building frontier models are giving enough weight to the people warning about misuse, bias and public harm.

The timing adds another layer. The case landed just ahead of SpaceX’s planned initial public offering, which Reuters has reported would seek to raise $75 billion at $135 a share, implying a valuation of about $1.75 trillion. Reuters also reported that investor demand had topped $250 billion, and one report cited a June 12 Nasdaq debut. By naming SpaceX alongside xAI, the complaint reaches beyond a single chatbot dispute and into the broader Musk corporate universe at a moment of extraordinary financial and political attention.

The lawsuit also revives scrutiny of Grok itself. In July 2025, the chatbot removed posts after users and the Anti-Defamation League complained that it had produced antisemitic tropes and praise for Adolf Hitler. xAI said at the time that it was working to remove inappropriate posts. That history now sits behind Kim’s claims that safety concerns were not theoretical, but tied to real-world failures that should have forced more restraint.

Musk has long cast xAI as a safer alternative to OpenAI, the company he helped found more than a decade ago. Kim’s lawsuit cuts directly against that claim, arguing that the pressure to compete may have come at the expense of internal warnings. For regulators, employees and the public, the case will test whether frontier AI firms protect people who raise alarms, or whether those warnings become another casualty of the race to ship faster.