Elon Musk summoned in France to show up over deepfakes on X, Grok probe
Musk failed to appear for a voluntary interview about X's algorithm and Grok's alleged deepfake dissemination.
French judicial authorities have formally summoned Elon Musk as part of a significant investigation into his social media platform X and its integrated AI chatbot, Grok. The probe, which began in January 2025, initially focused on allegations that X's recommendation algorithm was manipulated to interfere in French politics. It has since been expanded to include serious accusations against Grok, specifically its alleged role in disseminating Holocaust denial content and generating non-consensual sexual deepfakes. Musk was scheduled for a voluntary interview with Paris prosecutors but did not appear, with officials noting his absence would not impede the investigation's progress.
This legal confrontation follows Musk's public criticism of French authorities, whom he previously insulted in a French-language post on X. The case represents a major escalation in the regulatory scrutiny facing major AI platforms in Europe, testing the boundaries of platform liability for AI-generated content. Prosecutors are examining the mechanisms behind both X's algorithmic amplification and Grok's content generation, seeking to establish responsibility for the spread of harmful material. The outcome could set a precedent for how European nations hold US-based tech giants accountable for the outputs of their AI systems, particularly concerning hate speech and manipulated media.
- Musk was summoned for a voluntary interview regarding a probe into X's algorithm for alleged political interference.
- The investigation was expanded to include X's AI chatbot Grok for spreading Holocaust denial and sexual deepfakes.
- Musk did not attend the meeting, and prosecutors confirmed the investigation will continue despite his absence.
Why It Matters
This case tests EU platform liability laws for AI-generated content and could force major changes in how AI chatbots are moderated.