Former Engineer Files Lawsuit: Was He Forced to Leave xAI Due to Safety Concerns?
Devin Kim joined the AI startup in 2024 but claims he was fired for raising safety concerns about the chatbot Grok. His lawsuit alleges retaliation and wrongful termination.
- Home
- News
- Software & Development
Don't miss any news!
Devin Kim joined the AI startup xAI in 2024 as one of its first employees. However, his efforts to make the chatbot Grok safer were allegedly stifled, according to his lawsuit.
Former Employee Raises Allegations
According to the lawsuit, Devin Kim was one of the first employees at the AI startup xAI in 2024. Shortly after joining, he was promoted to a significant leadership position. As reported by The Guardian, Elon Musk allegedly expected appropriate safety measures to be implemented. However, his supervisor, xAI co-founder Jimmy Ba, reportedly ignored these instructions and dismissed Kim's demands for safety measures. In September of the previous year, the engineer was suddenly terminated—reportedly just before he was to present on AI safety to the company’s management.
The lawsuit states: "Mr. Kim repeatedly complained that xAI's failure to prioritize AI safety, particularly concerning Grok, practically guaranteed that the company would engage in unlawful acts, ranging from incitement to discrimination to the dissemination of weapons of mass destruction." The lawsuit accuses xAI, which is now part of SpaceX, of retaliation and wrongful termination in violation of California law, seeking unspecified damages. Both companies have not yet responded to requests for comment, according to The Guardian.
Deepfakes Have Consequences
On the day the lawsuit was filed, a Canadian regulatory authority also noted that Grok's image generation tool violated the country's privacy laws by allowing users to create non-consensual, sexualized deepfakes. This feature had sparked a global scandal earlier in the year. According to a study by the Center for Countering Digital Hate, approximately three million sexualized images were generated in just eleven days, including around 23,000 depictions of children.
Following a formal investigation by the Canadian Privacy Commissioner, xAI announced adjustments. Malaysia and Indonesia even completely banned the chatbot in response to the mass deepfake generation. The AI function is no longer available. In Germany, the scandal is expected to have repercussions: Federal Justice Minister Stefanie Hubig has introduced a bill that would criminalize the creation and distribution of intimate content as well as the violation of personal rights through convincingly realistic deepfakes.
Kim Continues to Advocate for AI Safety
As reported by Techcrunch, Kim's commitment to AI safety did not begin during his time at xAI. While working for Scale AI from 2022 to 2023, he led safety initiatives, including a project to create training data that would teach AI systems to recognize harmful content. Additionally, the nonprofit Center for AI Safety recently appointed Kim as its president.