Providers — companies such as OpenAI that build AI systems — must design those systems so that users are explicitly notified when they are interacting with AI rather than a human. They must also ensure that content generated or manipulated by their systems contains machine-readable marks so it can be detected as AI-created, according to the European Commission.
Deployers — the platforms and entities that use those systems — must clearly label deepfake videos, images and audio files, as well as AI-generated or manipulated text on matters of public interest. Under the rules, only deepfakes involving real and existing people, places, events or objects require labelling, Global News reported.
Deployers of AI-based emotion recognition and biometric categorization tools must also clearly inform users they are interacting with AI, according to the commission.
Optional icons to indicate AI
The European Commission has introduced a set of optional icons that platforms can use to indicate how much AI was involved in creating content. A fully AI-generated video would carry an “AI Generated” icon, while an authentic photograph with an AI-swapped face would require an “AI Modified” label, Global News explained.
Some platforms including TikTok and Instagram have already begun applying their own AI labels, and the EU rules make clear those would qualify — though the commission noted the labelling requirements themselves are not optional, even if the specific icons are.