eu: ai content gets marked from 2 december, sexual deepfakes of real people get banned

what happens on 2 december

two things, per article50.io's breakdown of the eu rules.

first, providers of generative ai systems that were already on the market before 2 august 2026 must mark their synthetic audio, images, video and text in a machine-readable way. newer systems have had to do this since august.

second, two new bans start: on systems that generate or manipulate non-consensual sexual deepfakes of identifiable people, and on child sexual abuse material.

what marking means

the law says outputs must be "marked in a machine-readable format and detectable as artificially generated or manipulated."

the technical solutions have to be "effective, interoperable, robust and reliable as far as this is technically feasible."

robust is the key word. it's meant to survive edits, not sit in a metadata field you can strip.

the deepfake ban

placing on the market or using a system that generates realistic sexual images, video or audio of an identifiable person without their explicit consent is banned from 2 december.

why it matters for a creator business

platforms are about to get a much easier way to tell what's generated.

and any content that looks like a real, identifiable person without consent is now squarely illegal in the eu.