A former engineer at Elon Musk’s AI company xAI has filed a lawsuit in a California state court claiming he was dismissed after raising repeated concerns about the safety of Grok, xAI’s AI chatbot. Devin Kim, who departed the company in September 2025, filed the complaint on Tuesday against both xAI and its parent company SpaceX. The lawsuit arrives days ahead of SpaceX’s expected public listing, which analysts expect could value the company at approximately $1.77 trillion.
According to the complaint, reviewed by TechCrunch, Kim was an early member of xAI’s post-training team and later led research tooling. During his tenure, he became increasingly vocal about what he saw as the company’s failure to prioritise safety in Grok’s development. His specific concerns were on Grok’s potential to generate discriminatory content and provide information about weapons of mass destruction.
His concerns proved to be well-founded when Grok made headlines in July 2025 after the chatbot produced responses in which it likened itself to Adolf Hitler in what became widely known as the MechaHitler incident. xAI took the model offline and later attributed the behaviour to unintended interactions in the system. Months after Kim’s departure, Grok was used to generate nonconsensual sexual imagery on X, Musk’s social media platform.
The complaint makes a pointed choice in who it holds responsible. Rather than implicating Musk, Kim’s lawyers describe him as having directed xAI to comply with the law and implement appropriate safety and testing processes. The target is instead xAI co-founder Jimmy Ba, a machine-learning researcher who left the company earlier this year.
Ba is portrayed in the complaint as someone who actively resisted safety measures and retaliated against Kim for raising them, in an effort to suppress his concerns. The lawsuit alleges Ba told Kim at one point that “AI will kill us all anyway,” and describes him as driven primarily by a mission to make xAI the first company to reach superintelligence.
One of the more serious allegations involves the release of Grok Code 1. The complaint claims Ba attempted to bypass EU safety regulations during that release, misrepresented aspects of the model to avoid legally required testing and indicated he would rather release an unsafe product than a slower-performing one. The lawsuit alleges Musk ultimately had to intervene.
Kim says he had been preparing a presentation of his findings for the week of 15 September 2025. Instead, Ba called him into a meeting and told him they should “go their separate ways,” without providing a satisfactory reason.
Kim is seeking compensatory and punitive damages and a declaratory judgment that xAI and SpaceX’s conduct was unlawful. He frames his dismissal as retaliation against a whistleblower, arguing that xAI’s alleged disregard for safety violated areas including internet regulation, consumer protection and arms and explosives regulation.
Kim’s background in AI safety predates his time at xAI. Before joining the company, he worked at Scale AI where he led projects producing training data for systems designed to detect harmful content. He was named president of the nonprofit Center for AI Safety last week, days before the lawsuit was filed.
The lawsuit lands at a moment when AI companies are under growing pressure from regulators, researchers and governments over how they manage safety before deploying increasingly powerful systems. The tension Kim describes at xAI, between moving fast to stay competitive and taking time to properly evaluate risks, is not unique to one company.
Whether his specific claims are proven in court, the case is likely to draw attention from regulators and investors alike, particularly on the ongoing scrutiny of how AI companies handle internal dissent.
Also Read: Microsoft reportedly preparing fresh Xbox layoffs amid strategic reset