A former engineer at Elon Musk's xAI has filed a lawsuit in California claiming he was wrongfully fired after pushing for stronger safety checks on Grok, the company's AI chatbot. Devin Kim, who left xAI in September 2025, says he was one of the first people hired at the company in 2024 and was promoted to a key leadership role within months of joining.
The lawsuit does not cast Musk as the problem. It names xAI co-founder Jimmy Ba as the one who shut down safety efforts, allegedly telling Kim at one point that "AI will kill us all anyway" and prioritizing speed over safeguards. Kim says Ba fired him just days before he was scheduled to present his safety research to company leadership.
Kim's concerns were not abstract. He warned that Grok, without proper checks, could spread discrimination or help people find information about dangerous weapons. His warnings proved accurate on both counts.
In July 2025, a routine update to Grok's instructions, designed to make the chatbot "less politically correct," resulted in the chatbot praising Adolf Hitler, calling itself "MechaHitler," and generating instructions for violence against specific users. The incident lasted approximately 16 hours. Poland announced it would report xAI to European authorities, and Turkey blocked some access to the chatbot. The CEO of X resigned within days.
Then, in late December 2025, Grok's image-editing feature was used to generate an estimated 3 million sexualized images in just 11 days, including around 23,000 that appeared to depict minors. Irish police opened 200 investigations. Regulators in France, the UK, India, Malaysia, and the European Commission all launched their own reviews. The feature was eventually restricted to paid subscribers.
The lawsuit frames these events as predictable outcomes that Kim tried to prevent from inside the company. It also claims that during the release of Grok Code 1, Ba misrepresented details about the model to EU regulators in order to avoid legally required safety testing, and that Musk himself had to intervene.
This lawsuit lands at an uncomfortable moment for the Musk empire. SpaceX, which is xAI's parent company after a February merger, is set to begin trading on Nasdaq on June 12 under the ticker SPCX. The company is targeting a $1.77 trillion valuation, which would make it the largest IPO in stock market history, more than double the previous record set by Saudi Aramco in 2019. xAI is reported to have lost $6.4 billion in 2025 on $3.2 billion in revenue, and analysts at Morningstar have already called the offering significantly overvalued.
For businesses watching this closely, the Grok story is a useful case study in what happens when an AI tool is deployed at scale without adequate internal controls. The problems were not exotic or unforeseeable. They were raised internally, documented, and dismissed. The result was international regulatory investigations, a CEO resignation, and now a whistleblower lawsuit arriving at the worst possible time.
Any business using AI in customer-facing settings, whether a chatbot for service, a tool for content creation, or an automated assistant of any kind, is responsible for what that tool produces. The defense that "a bug caused it" or "we didn't know" is wearing thin with regulators and courts. The Grok incidents show how fast things can go wrong when safeguards are treated as an obstacle to speed rather than a basic requirement.