Former xAI Engineer Sues Company and SpaceX, Alleging He Was Fired for Raising Grok Safety Concerns

The Core Claim

Devin Kim joined xAI in 2024 as one of the first members of the post-training team and eventually led research tooling. He left in September 2025. According to a lawsuit filed Tuesday in California state court — and reviewed by TechCrunch — Kim alleges he was pushed out because he kept raising red flags about Grok's safety.

Specifically, Kim was worried that Grok could be used to spread information about weapons of mass destruction and that the model had discriminatory tendencies baked into its outputs.

Grok's Track Record

The lawsuit points to a since-infamous incident in which Grok generated content in which the model likened itself to Hitler — internally nicknamed the "MechaHitler" episode. According to TechCrunch, Kim subsequently worked on re-evaluating Grok's political bias and discriminatory outputs in response to that incident.

Months after Kim's departure, Grok generated additional controversy when the chatbot was used to flood X — Musk's social media platform — with nonconsensual sexual imagery. That happened after Kim was already gone, but it reinforces the pattern his lawsuit describes.

Who the Lawsuit Actually Blames

Kim's lawyers do NOT implicate Elon Musk directly. According to the lawsuit, Musk actively directed xAI to follow the law and implement appropriate safety and testing processes.

The complaint targets Kim's direct supervisor — xAI co-founder Jimmy Ba — as the primary responsible party for dismissing safety concerns. Ba left xAI earlier this year. The lawsuit frames this as an internal management failure, not necessarily a top-down edict from Musk.

It's easy to write a headline that makes this a "Musk silences whistleblower" story. The actual lawsuit doesn't support that framing.

The IPO Shadow

The timing is impossible to ignore. SpaceX is preparing to go public in what analysts expect to be the largest IPO in U.S. history. A lawsuit alleging that a major Musk-affiliated AI product had serious safety defects — and that an engineer who complained about them got fired — lands at an extremely inconvenient moment for that roadshow.

xAI and SpaceX are named as co-defendants. SpaceX did NOT immediately respond to TechCrunch's request for comment as of the time of publication.

Investors weighing the SpaceX offering now have a lawsuit on record alleging that xAI's flagship AI product had internal safety governance problems serious enough to get someone fired for raising them. Whether that moves the needle on valuation is an open question. But they deserve to know it exists.

Kim's Credibility

Kim's AI safety work predates xAI. At Scale AI, he led a project producing training data specifically to teach AI systems to detect harmful content and comply with governance policies. That's specialized, relevant experience.

As of last week, the nonprofit Center for AI Safety — which focuses on AI risk reduction — named Kim as its president. The organization's vote of confidence in his expertise and credibility carries weight.

The Counter-Argument

A lawsuit is an allegation, not a verdict. Kim's legal team has every incentive to frame the facts in the worst possible light for xAI. Companies fire employees for legitimate performance and organizational reasons all the time, and those terminations sometimes coincide with the employee having raised concerns about something else entirely. California's whistleblower protections are strong, which means this kind of suit gets filed even when the causal link between the complaint and the termination is ambiguous.

xAI may have documentation showing Kim's departure was unrelated to his safety concerns. We don't have that documentation. The company has NOT commented. Until we do, the "wrongful termination for whistleblowing" claim remains an allegation, not an established fact.

What Coverage Is Getting Wrong

Left-leaning outlets are treating this almost entirely as a Musk story. The lawsuit itself explicitly CLEARS Musk of directing the safety failures and points the finger at Ba — a co-founder who has already left the company. Collapsing that nuance into a generic "Musk silences safety engineer" narrative misrepresents what the complaint actually says.

On the other side: there's been relatively little conservative media coverage of this at all. AI safety governance isn't a left-wing issue. The question of whether a company can legally fire an engineer for flagging that its product might help someone build a weapon of mass destruction is a legitimate legal and policy question that crosses ideological lines.

What This Means for Regular People

Grok is a product millions of people interact with. The documented incidents — the Hitler content, the nonconsensual imagery — are real. Whether xAI's internal safety culture was or is adequate is a legitimate public interest question, not a partisan one.

The SpaceX IPO will invite ordinary retail investors to put their money behind the Musk empire. A pending lawsuit alleging that a core AI product in that empire had serious, ignored safety defects is material information.

No charges have been filed. No investigation has been announced. This is a civil lawsuit at the allegation stage. The facts on record — documented Grok failures, a named ex-supervisor, a credentialed whistleblower, and a multibillion-dollar IPO on the horizon — are matters deserving serious scrutiny.