Just days before the largest initial public offering in history, Musk's technology empire was hit with an unwelcome lawsuit. Devin Kim, a former engineer at xAI, claims he lost his job because he repeatedly raised concerns about the safety risks of the Grok chatbot. He filed the lawsuit on Tuesday, targeting not only xAI but also its parent company, SpaceX.
Kim is not just any disgruntled employee. He joined xAI in 2024 as one of the company's first employees and rose to a leadership position within a few months. Today, he heads the nonprofit Center for AI Safety, an institute focused on the risks of artificial intelligence.
A dispute over how much safety the company can afford
What exactly was happening inside xAI? According to the lawsuit, Kim's work increasingly revolved around questions that the entire industry has yet to answer. How thoroughly should powerful chatbots be tested before being released to the public? How can harmful responses be limited without making the system less useful? And what level of risk is still acceptable for technology used by millions of people?
Kim claims management did not like his answers. The lawsuit states that he repeatedly warned that neglecting Grok's safety practically guaranteed that the company would sooner or later break the law. Among the risks, he cites the incitement of discrimination as well as assistance in the proliferation of weapons of mass destruction. The lawsuit also recalls a scandal that Grok actually experienced. During the so-called “MechaHitler" incident, the chatbot produced responses praising Adolf Hitler. The company apologized at the time and blamed the behavior on technical problems. Kim argues that stronger safeguards could have limited similar failures.
“AI will kill us all anyway"
A key figure in the dispute is xAI co-founder Jimmy Ba, a respected machine learning researcher and Kim's supervisor, who later left the company. According to the lawsuit, Ba rejected Kim's proposals for additional testing and safety checks, and became increasingly irritated by his insistence.
An interesting detail: Musk himself reportedly expected xAI to implement appropriate safety testing and processes. But according to the lawsuit, Ba ignored those instructions. And when Kim raised concerns, Ba allegedly responded with the sentence that put the lawsuit in headlines around the world: “AI will kill us all anyway." The context and meaning of the statement remain disputed, and Ba has not commented on it publicly.
Tensions came to a head last September. Kim says he was preparing to present his safety findings directly to the company's management. The presentation never took place. Instead, Ba summoned him to his office and told him that they should part ways.
Speed or safeguards?
Behind the personal dispute lies one of the most serious debates in the technology world. Artificial intelligence is developing at a dizzying pace. New models are released months apart, and companies compete fiercely for users, investment, and technological breakthroughs. Delaying a model's release means ceding the lead to competitors.
Meanwhile, governments and researchers are warning ever more loudly about misinformation, discrimination, and cybersecurity risks. Kim's dispute is a perfect example of this development. His lawsuit suggests that some members of management prioritized performance and deployment deadlines, while he insisted on stronger safeguards. This is quite ironic because Musk founded xAI in 2023 as a supposedly safer alternative to OpenAI, which he helped establish more than a decade ago. Last month, however, a jury rejected his own lawsuit against OpenAI, in which he claimed that the company had betrayed its original mission to benefit humanity.
This is not the first time safety concerns have been raised around Musk's companies. In 2023, Reuters documented more than 600 previously unreported workplace injuries at SpaceX, including crushed limbs, amputations, and one death. The company defends its record and points to its extensive safety training.
A lawsuit at a sensitive moment
Kim accuses xAI and SpaceX of retaliation and unlawful termination under California law and is seeking unspecified damages. Neither company has yet responded to a request for comment.
The timing plays an enormous role. SpaceX, which owns xAI, is set to go public on Friday, and according to available reports, the offering could value the company at approximately $1.77 trillion. It would be the largest initial public offering in history, and sources say investor demand exceeds the available shares several times over. That is precisely why one engineer's employment dispute has become a case watched by investors, regulators, and rival laboratories around the world.
Sources: techcrunch.com and timesofindia.indiatimes.com
