Paris — French prosecutors said investigators conducted a search of X’s French offices on Tuesday as part of an inquiry opened in January 2025 into alleged algorithm abuse and related harms. The Paris prosecutor’s cybercrime unit led the action, and Europol assisted, officials said. Elon Musk, X’s chairman, and former chief executive Linda Yaccarino have been summoned to appear at hearings in April; several X employees will be questioned as witnesses. The probe has been widened to examine sexually explicit deepfakes produced by X’s AI chatbot Grok, and X has blocked users from generating images of people in revealing clothing.
Key takeaways
- The search took place at X’s French offices; the underlying investigation was opened in January 2025 by the Paris prosecutor’s cybercrime unit.
- Europol was involved in the action, indicating cross‑border investigative cooperation.
- Elon Musk and Linda Yaccarino were summoned to hearings scheduled for April 2026; X staff will appear as witnesses.
- Prosecutors broadened the probe to include “Grok’s sexualized deepfakes,” per the Paris prosecutor.
- Last week the European Commission launched its own inquiry into Grok’s image‑generation capabilities amid complaints about sexually explicit outputs.
- X has stopped Grok users from creating images of people in revealing clothing while regulators review the system.
Background
AI chatbots and image‑generation tools have surged in capability and public use since late 2023, prompting regulators to revisit rules that previously focused on classical content moderation. Platforms that combine conversational models and image synthesis—like Grok—raise new legal questions because they can generate realistic but fabricated images on demand. European regulators have been especially active: the EU’s Digital Services Act and other initiatives have increased scrutiny of algorithmic amplification, transparency, and automated content generation.
French authorities opened the current inquiry in January 2025 reportedly on concerns about potential misuse of platform algorithms. The cybercrime unit of the Paris prosecutor’s office leads investigations that intersect online harms, child protection, and large‑scale automated content. Stakeholders now include national prosecutors, Europol, the European Commission, platform leadership at X, and user communities affected by generated content.
Main event
On Tuesday the Paris prosecutor’s office confirmed that its cybercrime unit carried out a search at X’s France offices; officials said Europol took part in the operation. According to the prosecutor’s statement, the probe—opened in January 2025—initially examined suspected misuse of recommendation or ranking algorithms and has since been widened to cover sexually explicit deepfakes created by Grok.
Prosecutors have summoned Elon Musk and Linda Yaccarino to appear at hearings in April 2026, and they said multiple X employees will be questioned as witnesses. The summonses and witness interviews indicate prosecutors are seeking both leadership-level explanations of policies and technical detail from people who work directly on Grok and related systems.
European regulators have separately announced a probe into Grok’s image‑generation feature, following reports that the chatbot could be prompted to produce nude or sexually explicit images, including those of women and reportedly of children. In response, X disabled the feature that allowed Grok users to create images of people in revealing clothing while the platform and regulators assess the issue.
Analysis & implications
The raid and parallel European Commission inquiry mark a notable escalation in how regulators and law enforcement treat AI‑driven content risks. Legal exposure for X could include fines under EU digital rules, administrative orders to alter algorithmic behavior, and criminal inquiries if laws protecting minors or prohibiting certain kinds of imagery are implicated. Cross‑border cooperation via Europol suggests investigators are pursuing evidence that spans jurisdictions and server locations.
For platform governance, the incident underscores tensions between innovation, user demand, and legal compliance. Firms running multimodal AI systems now face the dual task of building safer defaults while preserving functionality; failure to do so risks regulatory penalties and reputational damage that can affect user engagement and advertiser relationships. Investors and partners may press for faster, more transparent mitigation measures.
Operationally, summonses for senior executives signal a pivot toward accountability at the leadership level. Even if prosecutions do not follow, hearings and regulatory inquiries can force public disclosures, remedial audits, or structural changes in how training data, prompts, and content filters are managed. Globally, other jurisdictions are likely to watch the outcome closely and may replicate investigative or regulatory steps.
Comparison & data
| Aspect | January 2025 Opening | February 2026 Expansion |
|---|---|---|
| Primary focus | Suspected algorithm abuse | Includes Grok sexualized deepfakes |
| Actors involved | Paris cybercrime unit | Paris cybercrime unit, Europol, European Commission (probe) |
| Platform response | Under review | X blocked users from creating revealing‑clothing images |
The table above shows how the inquiry has evolved from an algorithmic‑abuse investigation into one that explicitly covers AI image synthesis outputs. That shift reflects broader regulatory attention: where initial probes targeted recommendation and moderation systems, investigators are now incorporating risks posed by generative models that produce realistic but fabricated media.
Reactions & quotes
Officials and institutions have framed the action in terms of investigative scope and regulatory oversight.
“Grok’s sexualized deepfakes”
Laure Beccuau, Paris prosecutor (statement reported)
“We have opened a probe into the chatbot’s image‑generation features,”
European Commission (press office, reported)
“X has suspended generation of images of people in revealing clothing while the issues are assessed,”
X (company action reported)
Each quoted fragment above is brief and reported in official or public statements; prosecutors and regulators are emphasizing the investigative and precautionary steps rather than announcing charges at this stage.
Unconfirmed
- Public reporting has not quantified how many explicit images were generated or whether verified minors were involved; those details have not been publicly confirmed.
- It is not yet public whether the April hearings will lead to formal charges against executives; prosecutors have only announced summonses and witness interviews.
- The precise technical changes X will make to Grok or the timeline for restoring any suspended features remain unspecified.
Bottom line
The Paris raid and the European Commission’s probe signal that AI‑driven image generation is now a central target for both law enforcement and regulators in Europe. For X, the immediate priorities are legal cooperation, internal review, and transparent remediation to limit harm and demonstrate compliance; failure to act could trigger fines, ordered changes, or wider reputational impacts.
Watch for the April hearings and any follow‑up filings from the Paris prosecutor for concrete next steps; separately, the European Commission’s inquiry may produce policy or enforcement actions that affect other platforms and providers across the EU. The episode illustrates how fast‑moving AI capabilities are prompting faster, more coordinated responses from regulators and investigators worldwide.