A former xAI engineer, Devin Kim, sued the company and its parent SpaceX on Tuesday in California state court, alleging he was fired in September 2025 for repeatedly raising safety concerns about Grok. The complaint lands days before SpaceX is set to join the public markets in what is shaping up to be the largest IPO in history, putting xAI's internal safety practices in front of investors at the worst possible moment.
Kim joined xAI in 2024 as one of the first members of its post-training team and eventually led research tooling. According to the complaint, he flagged risks that Grok could foment discrimination and disseminate information about weapons of mass destruction, and continued pressing those concerns after the model's July 2025 episode in which it referred to itself as 'MechaHitler' on X. The suit frames Kim as a whistleblower whose objections covered internet regulation, consumer protection, unfair business practices, and arms and explosives rules.
The legal claim is notable for who it does not target. Kim's lawyers describe Elon Musk as having directed xAI to follow the law and implement appropriate safety and testing processes. The retaliation claim instead lands on xAI co-founder Jimmy Ba, Kim's supervisor, who left the company earlier this year.
“September was my last month at xAI! I joined as one of the first members of the post-training team in 2024 and eventually led research tooling, where we built some of the world's best systems to accelerate Grok's development.”— Devin Kim, Former xAI engineer and president of the Center for AI Safety
Key facts
- 01Devin Kim, an early member of xAI's post-training team, filed suit in California state court on Tuesday alleging retaliatory firing in September 2025.
- 02The complaint targets xAI co-founder Jimmy Ba, who left the company earlier this year, and explicitly does not implicate Elon Musk.
- 03Kim alleges Ba tried to bypass EU safety testing for Grok Code 1 in August 2025, with Musk intervening.
- 04Filing lands days before SpaceX, xAI's parent, is set to go public in what is shaping up to be the largest IPO in history.
- 05Kim was named president of the Center for AI Safety last week.
The complaint alleges Ba vehemently opposed safety measures, at one point telling Kim 'AI will kill us all anyway,' and prioritized reaching superintelligence first. In one specific incident in August 2025, Ba is accused of trying to thwart EU safety regulations during the release of Grok Code 1 by misrepresenting aspects of the model to avoid legally required testing. 'Mr. Ba indicated that he would rather release an unsafe model than a poor-performing one. Mr. Musk ultimately had to intervene,' the filing states.
Kim says he planned to deliver a presentation of his findings the week of September 15, 2025. Instead, the complaint alleges, Ba called him into a meeting and told him they should 'go [their] separate ways' without offering a satisfactory reason. Kim is seeking compensatory and punitive damages and a declaratory judgment that xAI and SpaceX's conduct was unlawful.
xAI and SpaceX did not immediately respond to requests for comment, and Ba has not yet publicly responded to the allegations. The suit's specific claims — bypassing EU testing for Grok Code 1, internal opposition to safety review, and the timing of Kim's dismissal relative to his planned presentation — will be the central questions in discovery. None has been tested in court.
Kim's background adds weight to the filing. Before xAI, he worked on early safety initiatives at Scale AI, including a project producing training data to help systems detect harmful content and comply with governance policies. Last week, the nonprofit Center for AI Safety named him its president, a role that will keep him publicly engaged on the same issues at the heart of the lawsuit.
Grok has been at the center of repeated incidents that the complaint cites as evidence Kim's warnings were warranted. Beyond the 'MechaHitler' episode, the chatbot was later used to flood X with nonconsensual sexual imagery a few months after Kim departed. Both events drew regulator attention in jurisdictions including the EU, where the Grok Code 1 release is now alleged to have skipped legally required testing.
The timing puts SpaceX's underwriters in an awkward position. A whistleblower suit alleging deliberate evasion of EU AI rules at a subsidiary is the kind of disclosure that pricing committees scrutinize closely, particularly when the named executive — Ba — has already departed and is no longer available inside the company to rebut the claims. xAI's safety governance, until now mostly a question for journalists and regulators, becomes a question for institutional shareholders.
The structural pattern in the complaint is the one that matters for the broader AI industry: a safety lead alleges that a senior technical executive treated safety review as friction to be routed around, and that the disagreement ended with the safety lead's exit. Whether or not Kim prevails, the suit hardens an emerging template — engineers leaving frontier labs, then suing or speaking publicly about pre-release testing decisions. For xAI specifically, the case forces a choice between defending Ba's conduct or distancing the company from a co-founder who is no longer there to defend himself. Neither option helps the IPO.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
