Grok in Paris the legal battle for the IA that generates sexual images

The French police conducted a search at the X offices in Paris this week as part of a criminal investigation that has as its axis the artificial intelligence tool Grok, used by the platform to generate images with sexual content. The authorities opened the file in early 2025 and, according to official sources, extended investigations after receiving complaints about the generation and dissemination of illicit material, including sexual deepfakes and content that deny the Holocaust.

The intervention was carried out by the unit specialized in cybercrime of the National Gendarmerie and was supported by Europol agents. The Paris prosecutor has promoted measures that go beyond the register: he voluntarily quoted Elon Musk and X's executive director, Linda Yaccarino, to come to testify on April 20, and plans to receive several employees from the platform as witnesses between April 20 and 24. The public prosecutor's office justified these subpoenas by stating that the statements will allow the company's officials to state their version of the facts and, where appropriate, the compliance measures they intend to take ( Public prosecutor's statement).

A number of possible crimes are at the centre of the investigation: complicity in the possession and distribution of child pornography, crimes related to unconsensual sexual deepfakes, Holocaust denial, fraudulent data extraction, system manipulation and the alleged exploitation of an online platform in an illegal manner within an organized structure. The list of charges reflects the complexity of the case: it is not just a matter of moderating content, but of determining criminal and technical responsibilities in the management and deployment of IA models that interact in public environments.

X has already publicly reacted to the investigation, or at least certain aspects of it, as a politically motivated action; its official account of Global Affairs published a version of the facts in which it rejected the allegations about algorithmic manipulation and "fraudulent data extraction" ( message from X Global Affairs). The tension between the company and regulators is not new: from Brussels to London and California, steps have been taken to assess whether the company complied with the legal obligations by launching Grok.

The European Commission launched an investigation in January 2026 to check whether X carried out an appropriate risk assessment required by the European regulatory framework known as Digital Services Act before deploying Grok; the DSA forces platforms to identify and mitigate systemic risks arising from their services ( more about the DSA). In parallel, national agencies have launched investigations of different scope: the Office of the United Kingdom Information Commissioner issued a statement expressing concern, Ofcom announced an investigation focused on online security and the California Attorney General opened a review for the use of IA in the generation of sexual images without consent ( ICO, Ofcom, California Attorney General's Office).

These actions are in the midst of another European Union sanction: in December, the Commission imposed a fine on X for non-compliance with transparency obligations associated with the DSA. This background underlines that the European authorities are vigilant about how the major platforms document and communicate the risks and controls of their automated systems, especially when they impact fundamental rights and public security.

What does this case raise from technology and public policy? First, it exposes the difficulty of containing the misuse of generative models when combined with a public interface and with the possibility of viral content sharing. Secondly, he stressed the need for clear frameworks for auditing and monitoring models: to know what data were used to train them, how requests that could lead to images that violated the law were filtered, and what mechanisms existed to curb the distribution of illicit material.

There are also implications in the field of technical and legal responsibility. Regulators seek answers on whether the platform implemented effective safeguards and whether there were failures or practices that facilitated mass data extraction or system manipulation. The line between moderation negligence and criminal conduct is not always clear, and there goes the work of prosecutors and computer experts to rebuild facts and attribute guilt.

For technology companies, the lesson is clear: launching generative models in powerful environments without robust controls can have severe legal and regulatory consequences. For regulators, the challenge is to act quickly but also with sound technical criteria, because imposing excessive restrictions or sanctions without evidence can feed political persecution narratives, while a warm response would leave people affected by deepfakes or sexual exploitation unprotected.

At the social level, the crisis revives questions about the protection of privacy and dignity from tools that reproduce images without consent. In addition, the presence of content that denies historical crimes adds a layer of symbolic and legal damage; many jurisdictions consider Holocaust denial as a crime, and its proliferation on platforms raises dilemmas about moderation, education and freedom of expression.

As instruction takes place in France and investigations in several countries are moving forward, two parallel processes should be observed: on the one hand, the criminal investigation that can end in charges and sanctions; on the other, the regulatory procedures that seek to adjust the governance of the IA in the public sphere. Both will influence how companies design and deploy generative models in the near future.

History is not reduced to a single social network or to a single product: is a test for the entire ecosystem on how to balance innovation and security, and on what models of multi-jurisdictional governance are effective against risks that do not respect borders. If you want to consult the official communications related to the case and the regulatory investigations, you can review the Paris Public Prosecutor's report ( ad in LinkedIn), the details of the prosecutor in PDF ( Communiqué of the Paris Court) and reactions from regulatory bodies such as the European Commission on DSA, ICO, Ofcom and California Attorney General's Office. You can also follow the official position spread by X through your Global Affairs account ( message in X).

In short, this case is a crossroads: how to allow technology to move forward and, at the same time, how to demand that such progress do not lead to impunity for harm to people or undermine criminal and living conditions. The next steps - audiences, experts and regulatory resolutions - will mark not only the course of X and Grok, but also the necessary rules for the IA that will live with millions of users worldwide.