EU Parliament Moves to Ban AI Nudifiers and Delay High-Risk Rules

Parliament Draws a New Red Line on AI

The European Parliament is voting this week to add AI nudification systems to the AI Act's list of outright prohibitions. The same vote extends the compliance deadline for high-risk AI provisions, giving companies and regulators more time to prepare for the Act's most demanding requirements. Both measures were debated during the March 25-26 Brussels plenary session, according to the Parliament's published agenda.

The nudification ban targets AI tools that generate non-consensual intimate imagery of real people, a category of harm that has exploded in prevalence since generative AI went mainstream. By adding it to the prohibited practices list rather than treating it as a high-risk application, Parliament signals that no amount of safeguards or transparency measures can make this technology acceptable.

More Time for High-Risk Compliance

The high-risk AI provisions were originally scheduled to take effect in August 2026. But implementation standards have fallen behind schedule, and the supporting tools that companies need to demonstrate compliance don't yet exist in final form. MEPs are adopting a position that once these standards and tools are available, organizations will have six months to comply, with the overall extension capped at December 2027.

This isn't a retreat from regulation. It's an acknowledgment that rushing enforcement without ready infrastructure would produce chaos rather than accountability. The AI Act already applies its bans on social scoring and manipulative AI, which took effect in February 2025. The transparency obligations for general-purpose AI models kicked in earlier this year. High-risk requirements represent the final, most complex layer.

Why This Matters

For European companies building AI products, the extended timeline is a practical relief. For everyone else, the nudification ban is the headline. It puts the EU further ahead of any other jurisdiction in explicitly criminalizing AI-generated non-consensual imagery at the regulatory level. The United States has no federal equivalent. The UK's Online Safety Act addresses distribution but not generation.

Companies operating in the EU should use the extra compliance runway wisely. The deadline will arrive, and when it does, the bar for high-risk AI systems, from hiring algorithms to medical diagnostics, will be the highest anywhere in the world.

Sources

• European Parliament Plenary Newsletter, March 25-26, 2026