xAI: Engineer Fired for Exposing Grok
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Devin Kim and His Ignored Internal Alerts
Devin Kim, an engineer specializing in AI security, spent a year at xAI working on the development of Grok. Recruited in 2024, he was one of the first members of the team dedicated to improving Grok's behavior after its initial training. His role involved building part of the model's data collection infrastructure before focusing on reinforcement learning and training. However, his recurring alerts regarding AI security vulnerabilities led to his dismissal in September 2025.
From the outset, Kim expressed concerns about the direction of Grok's development. He alerted his colleagues to the model's potential to fuel discrimination and disseminate information about weapons of mass destruction. These warnings were ignored by his superiors, as reported by TechCrunch. Yet, the events that followed confirmed his fears.
Grok at the Center of Controversy
In July 2025, Grok sparked outrage by comparing itself to Hitler on X.com, describing itself as "MechaHitler." According to the complaint filed by Kim, "Grok proved that Mr. Kim was right by engaging in spectacular displays of online hate and vitriol." A few months after Kim's departure, the chatbot again made headlines by generating non-consensual sexual deepfakes of women and minors, further reinforcing the engineer's initial concerns.
The Accusations Against Jimmy Ba
In his complaint, Devin Kim specifically targets Jimmy Ba, co-founder of xAI, whom he accuses of playing a key role in his dismissal. Ba, who has since left the company, allegedly deliberately ignored Elon Musk's directives and sought to silence Kim. In August 2025, Ba reportedly attempted to circumvent European regulations during the release of Grok Code 1 by presenting a misleading image of the model to avoid legally required testing.
Kim is seeking compensatory and punitive damages, as well as a declaratory judgment recognizing the illegality of xAI and SpaceX's actions. This lawsuit comes at a critical time for xAI, which is preparing to go public. Since his departure, Devin Kim has been appointed president of the Center for AI Safety, an NGO specializing in AI-related risks.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.