EU AI Act enforcement: Deepfake ban targets parliament members

On 5 October, Agora Digitale Transformation published an investigation into non-consensual deepfake pornography targeting European lawmakers. The report identifies 50 members of the European Parliament as victims, including 46 women. This equates to roughly one in six female MEPs being affected.

Politico reported on 5 October that the EU is currently investigating the use of Grok on the social media platform X. This investigation runs alongside a ban on AI-powered nudifier apps scheduled for 2 December. The urgency of enforcement has shifted from a legal to a political priority, particularly as high-profile figures become victims.

On 30 September, a cross-party group of 119 MEPs sent a letter to Commission President Ursula von der Leyen and other EU leadership figures. They demanded effective enforcement of the ban and a prompt resolution of the Grok investigation. "There can't be any loopholes or delays," said Lena Schilling, an Austrian Greens MEP and signatory, according to Politico.

The enforcement challenge

Benjamin Shultz, a co-author of the report, noted that politicians cannot simply opt out of having their likenesses online, making them vulnerable targets for AI nudifiers. He argued that the Commission must hold all entities accountable, including payment processors and domain registrars that inadvertently enable the industry. Shultz also suggested the Commission look into de-indexing AI nudifier websites from search engines to reduce their visibility. The technical reality of enforcing these rules is complex because the infrastructure supporting these apps is decentralized and often operated by entities outside direct EU jurisdiction.

EU regulators are moving from drafting frameworks to active policing. The AI Act is often criticized as slow and bureaucratic, but the deepfake scandal has forced a faster response. The EU is using its existing regulatory power to pressure platforms to act before the formal ban date. This is a test case for how Brussels will handle other AI-related violations, particularly concerning generative models that are difficult to audit. The outcome will influence how future compliance mechanisms are designed and enforced across member states.

On 7 October, Tech.eu reported that DIGITALEUROPE's director-general, Cecilia Bonefeld-Dahl, argued that Europe cannot regulate its way to tech leadership. She emphasized the need for faster access to funding and less regulatory complexity, suggesting that the current pace of enforcement might be stifling innovation in areas beyond AI safety. This creates a tension between the need for strict AI governance and the desire for a competitive tech ecosystem. Industry leaders often argue that excessive compliance costs drive innovation to other jurisdictions, a claim that regulators dispute by pointing to the consumer protection benefits of strict rules.

Broader AI governance context

The focus on AI regulation in Europe is not limited to deepfakes. On 8 October, ZDNET reported on California's "No Robo Bosses Act," a law that protects workers from fully automated personnel decisions. While this is a US state law, it signals a global trend of regulators intervening in AI's role in the workplace. The EU is watching these developments closely. The AI Act also includes provisions for high-risk AI systems, including those used in employment and education. The convergence of these regulatory efforts suggests a global shift toward stricter oversight of automated decision-making processes in sensitive sectors.

In the US, Peter Thiel's recent lectures in Nashville have drawn attention to the political dimensions of AI regulation. On 9 October, Gizmodo reported that Thiel has framed AI regulation as an existential threat. He has significant financial stakes in AI companies and linked the issue to theological concepts. This rhetoric reflects a broader resistance to AI governance in parts of the US tech industry, contrasting with the EU's more regulatory approach. The EU's actions on deepfakes may serve as a counter-narrative to this resistance, demonstrating that the bloc is willing to enforce its rules even when they are politically unpopular with industry players who benefit from looser oversight.

As the 2 December ban approaches, the Commission faces a critical test. The ability to enforce the AI Act against powerful platforms will determine its credibility. The targeting of 50 MEPs has raised the stakes, turning a technical enforcement issue into a matter of political security. The outcome of the Grok investigation and the implementation of the nudifier ban will set the tone for AI governance in Europe for years to come. Regulators are no longer just writing rules; they are actively trying to protect citizens from the harms of unchecked AI systems that operate with minimal human oversight.