EU requires firms to label 'made with artificial intelligence' as new deepfake rules kick in

The European Union has begun enforcing new transparency obligations for companies that create or use content generated by artificial intelligence, including so-called deepfakes, images and text intended for public consumption.

Under the EU's comprehensive AI law, companies must make it clear to users when they are interacting with AI systems such as chatbots, and must label images or texts produced by AI so that audiences can recognise synthetic material.

How the rules work

Compliance can be achieved through digital watermarks and other visible markers designed to make AI-generated material easy to spot, and firms that fail to follow the rules face substantial fines. The law applies to AI used for professional purposes; people using AI solely for personal tasks are exempt.

An EU representative warned of the risks the bloc aims to address: "Generative artificial intelligence makes possible the production of disinformation on a scale never seen before, tailored to specific audiences and spread at extraordinary speed." The representative added the goal of the rules is "to preserve citizens' ability to trust what they see, hear and read."

Content produced to inform the public on matters of general interest must be labelled if it was created by AI without human editorial oversight. Existing AI systems have until 2 December 2026 to comply with the new obligations, while exemptions are carved out for "artistic, creative, satirical and fictional" works.

Industry reaction has been mixed. Some voices say the new duties will be hard to implement at scale and could lead to most online content carrying AI labels, while companies are already introducing their own tagging systems ahead of the law.

For example, TikTok has required creators for several years to mark AI-created images, audio and video and says more than three billion pieces of content already carry such labels through its tools and recognition technology. Meta has introduced an "AI Info" label on Instagram and Facebook. Google has signed the EU Code of Practice on AI transparency and says it is working with firms including Nvidia, OpenAI and Apple on digital labelling tools.

Industry experts also warn of implementation challenges. "We've heard that it will be very, very difficult to enforce. But I think that's often what we hear when it comes to meeting regulatory requirements. Still, life goes on and we find a way to resolve those issues," said Ashley Kasovan of the International Association of Privacy Professionals. Karen Massin of Google cautioned that "If internet content is flooded with overlapping AI labels and legal notices, people will find it harder to get the clear context they need."

As a final note, the EU has set aside 282ac10 billion to support construction of AI gigafactories.

Photo: press material from the event