A political agreement on March 11 for amendments to the Artificial Intelligence (AI Act) Law will add an explicit ban on non-consensual deepfake images generated by AI, as a direct result of the Grok scandal.
See also: Meta Oversight Board: Deepfakes detection inadequate
It took a scandal, a wave of regulatory outrage, and an alliance of 57 members of the European Parliament to achieve this, but the landmark EU Artificial Intelligence Act will now include an explicit ban on non-consensual sexual deepfakes.
On March 11, 2026, the MEPs reached a political agreement on a package of amendments to the EU AI law, with the ban on non-consensual images generated by AI, including child sexual abuse material, being highlighted as one of the most controversial and significant elements of the agreement.
The agreement, which was brokered between centre-right and centre-left MEPs in the European Parliament, also includes looser compliance rules for AI systems that are embedded in sector-regulated products, such as medical devices and industrial machinery.
The package is part of the broader AI Act Omnibus, a set of amendments being negotiated to simplify and, in some cases, strengthen the 2024 law as it moves toward full implementation.
See also: Deepfake attacks and biometric spoofing
The ban on sexual deepfakes was not originally planned to be part of the Omnibus. It was introduced in the wake of one of the most high-profile AI controversies to hit European regulators recently: the Grok case.
In late December 2025, Elon Musk's AI company, xAI, updated the Grok chatbot, integrated into the X social networking platform, with a new image editing capability. Within a few days, users were exploiting it to create realistic sexual images of real women and girls without their consent, including content that regulators said depicted minors in a way that constituted child sexual abuse material.
Between January 5 and 6 alone, researchers at the Paris-based nonprofit AI Forensics estimated that at least 6,700 sexual images were created through the tool.
The European Commission reacted quickly. The Commission's digital affairs spokesperson publicly described the content as ‘abominable’ and ‘clearly illegal’, stating that ‘it has no place in Europe’.
The European Commission confirmed on March 11 that the existing EU law, including the AI Act as written, did not prohibit AI systems capable of generating material of child sexual abuse or sexually explicit deepfake nudity. This legal gap, which was publicly recognized by the Commission in a letter to an MEP, provided the political impetus for amending the Omnibus.
See also: United Kingdom: Law criminalizing Grok deepfakes
France and Spain had supported the explicit ban throughout the Omnibus negotiations, with Germany and Slovakia joining them by threatening to block the file if it was not included. The EU Council, which represents the member states, added the ban in their place at a last‑minute adjustment.
