EU rules require clearer labelling of AI-generated content

Sunday 2nd August 2026 on 20:45 in Denmark Denmark

artificial intelligence, denmark, European Union

New European Union rules on artificial intelligence enter into force today, requiring clearer labelling of AI-generated content and making it easier for people to recognise when they are interacting with an AI system, DR reports.

Companies operating chatbots and other AI systems must clearly state that users are not communicating with a human. Deepfakes, including AI-generated images, videos and audio, must also be labelled as AI content.

The rules apply to public authorities, companies, institutions and private individuals that publish AI-generated content.

Morten Jensen, head of office at the Danish Digitalisation Agency, which is responsible for the EU regulation on AI, said the rules should make AI-generated content easier to identify.

“Citizens can increasingly expect content they encounter online that has been generated or manipulated using AI to become easier to recognise,” he told Politiken.

The rules are intended to increase transparency online, reduce manipulation using AI and help people make informed decisions about whether content can be trusted. Some AI systems must also include machine-readable labels, allowing technical systems as well as people to identify AI-generated material.

One of the most concrete requirements concerns interactions with chatbots. When customers use a chatbot on a company website, for example to help choose a product, it must be clear that they are speaking to an AI system rather than a person.

The rules apply both to providers of AI systems and to organisations that use them when communicating with customers. Dansk Industri supports rules that could strengthen trust in AI and create more uniform conditions across the EU.

Bo Wiberg, team leader for digitalisation policy at Dansk Industri, said all companies should assess how they use AI systems. The requirements will be particularly relevant to companies developing AI solutions, offering AI-based services or using AI to generate content aimed at customers, citizens or the public, he said.

The European Commission said in a press release that the rules will only be effective in individual member states if the appropriate authorities are appointed to supervise them. It is not yet known which authorities will oversee the rules in Denmark.

Source 
(via DR)