Europe Blinks — But Draws a New Red Line: No AI Nudification. Ever.
After nine hours of negotiations in Brussels, the European Parliament and the Council of the EU reached a provisional agreement on May 7, 2026 that will reshape the EU AI Act in two directions simultaneously — relaxing it for business, and tightening it where it matters most.
The retreat: High-risk rules delayed to December 2027
Rules for high-risk AI systems — including biometrics, critical infrastructure, education, employment, migration, asylum and border control — are now postponed by a year, set to apply from December 2, 2027. This is a significant concession to industry. Businesses had complained loudly that the original August 2026 deadline was unworkable, and Brussels listened.
Critics say the deal shows Europe caving to Big Tech. The agreement came after months of pressure from tech companies and — notably — the Trump administration, which has made no secret of its view that European AI regulation is economic protectionism dressed up as ethics.
But the story does not end there.
The advance: A complete ban on AI nudification
The Digital Omnibus prohibits AI systems that generate non-consensual sexually explicit content and child sexual abuse material — including so-called "nudification apps" that digitally remove people's clothing. The ban covers explicit images, videos or audio created without a person's consent. Companies will have until December 2, 2026 to bring their systems in line.
The decision to ban AI that creates sexually explicit images of individuals without their consent greenlights a provision first proposed after Elon Musk's Grok chatbot began publishing millions of nudified images of unwitting victims. The European Parliament pushed this through against initial resistance. It is one of the clearest examples of European regulation responding directly to a specific, named harm caused by a specific, named AI system.
The fine print that matters most
The deal reinstates the requirement for AI providers to register their systems in the EU database — even where they consider their system exempt from high-risk classification. This closes a potential loophole. Companies cannot simply self-declare their AI as low-risk and walk away. They must file that position publicly and stand behind it.
The law also delays the application of watermarking obligations on AI-generated content until December 2, 2026. Every piece of AI-generated content — text, image, video, audio — will need to be identifiable as such. For GAFAM companies operating in Europe, this is not a minor compliance checkbox. It is a fundamental change to how AI-generated content is published and distributed.
What this means for GAFAM
Google, Apple, Meta, Amazon and Microsoft all have AI systems operating in Europe that fall into high-risk categories. The delay to December 2027 gives them breathing room — but the EU AI Act has been delayed once. It will not be delayed twice. The new dates are binding. The clock is running. The nudification ban, meanwhile, lands directly on Meta — whose platforms have already been cited in European investigations — and on any company whose AI tools can generate or distribute non-consensual explicit content. December 2, 2026 is not far away.
The European perspective: Brussels blinked on the timeline. But it held firm on human dignity. That is not a bad trade — if the December 2027 deadline actually holds. gafam.ai will be watching.