Elon Musk skips Paris prosecutors' summons over illegal Grok images

Elon Musk did not attend a voluntary interview with Paris prosecutors, who are investigating Grok's role in generating prohibited sexualized images, including material involving minors. For xAI, this is no longer an isolated moderation scandal but a full-blown international legal case with political and regulatory consequences.

Why Musk was summoned

On April 20, 2026, the Paris prosecutor's office expected Musk for a voluntary interview as part of an investigation run by its cybercrime unit. He did not come, and the office simply recorded his absence. Former X chief Linda Yaccarino was also summoned to give explanations, and French investigators planned to question other company employees during the same week.

Formally, this is neither an arrest nor a formal charge, but the summons itself shows that the case has moved far beyond routine complaints about user-generated content.

The investigation was originally opened in January 2025 after complaints about X's algorithms and possible interference in French politics. In November 2025, however, its scope widened sharply: it now covers five possible criminal offenses. They include alleged complicity in the possession and distribution of pornographic material involving minors, the distribution of explicit deepfakes, tampering with an automated data processing system, and unlawful data extraction.

For xAI, this is a dangerous shift: the question is no longer how controversial a feature was, but whether the company may have allowed a systemic violation of the law.

What was found in Grok

The key episode concerns image generation in Grok. Investigators and plaintiffs in other countries claim the tool let users upload real photos of women and girls and then produce sexualized or nude versions of them without consent. That capability made the story toxic not only for X but for all of xAI, because it points not to random model failures but to a repeatable pattern of abuse.

• According to the Center for Countering Digital Hate, Grok created about 3 million sexualized images between December 29, 2025 and January 8, 2026.
• Researchers estimate that roughly 23,000 of them may have depicted minors.
• At its peak, the system generated up to 190 such images per minute.
• Up to 41% of all 4.6 million Grok images in that period contained sexualized images of women.

The problems did not start yesterday. On December 9, 2024, xAI launched the Aurora model but pulled it a few hours later, after it generated photorealistic images of real people without adequate restrictions. Then, on December 20, 2025, Musk announced that Grok could edit and create images directly inside X, and abuse grew sharply after that.

On January 9, 2026, xAI restricted image generation to paid subscribers, and on January 14 it said it had fully disabled photo "undressing" functions. But follow-up checks in February and March showed that workarounds and vulnerabilities had not gone away.

Why the dispute went global

Two days before the summons, on April 18, 2026, the US Department of Justice refused to help the French investigation. The American side said France was trying to use criminal law to regulate a public platform and free expression, an approach it said contradicts the First Amendment.

Paris responded, predictably, in firm terms: the prosecutor's office reminded that the French constitution guarantees the separation of powers and the independence of the judiciary. In other words, Paris signaled that it does not see this as a political attack and does not intend to drop the case because of Washington's position.

France is not acting alone. In January, Malaysia and Indonesia blocked Grok completely. The European Commission opened formal proceedings under the Digital Services Act and demanded that internal documents and technical data related to Grok be preserved until the end of 2026. A court in Amsterdam ordered xAI to stop generating non-consensual intimate images in the Netherlands under threat of daily fines. British regulators have launched parallel checks, and lawsuits and complaints have been filed in the US and Switzerland. In total, xAI now faces more than a dozen international proceedings.

What this means

The Grok affair shows that image generation is no longer just a flashy product feature. If a model lacks strict limits and a fast mechanism for blocking abuse, the debate quickly shifts from free expression to an investigation into harm, platform liability and the personal role of management.

The signal for AI companies is simple: releasing image tools without reliable guardrails is now not just risky but legally expensive and reputationally dangerous.