The AI Act transparency rules apply across the European Union from 2 August 2026, requiring users to be informed when they are interacting with artificial intelligence and introducing disclosure requirements for deepfakes and other synthetic content. The provisions directly affect companies, public bodies, media organisations and digital platforms operating in Denmark, Sweden, Finland and the rest of the EU.
The new obligations are contained in Article 50 of the AI Act, the European regulation establishing common rules for artificial intelligence. They are intended to make it easier for people to distinguish human communication and authentic material from content produced or manipulated by AI.
What the AI Act transparency rules require
Providers of chatbots and other systems designed to interact directly with people must ensure that users are told they are communicating with an AI system rather than a human, unless this is already evident from the circumstances.
The disclosure must be clear and distinguishable and must generally appear no later than the first interaction. The European Commission’s guidelines provide practical examples and clarify which companies are considered providers and which organisations are responsible for deploying a system.
Providers of generative AI must also add machine-readable markings to synthetic or manipulated text, images, audio and video. These technical markers are intended to help platforms, researchers and verification services detect the artificial origin of digital material.
The obligation is subject to limits. It does not generally cover systems used only for standard editing or tools that do not substantially alter the original content. Following amendments introduced through the EU’s AI simplification package, systems already placed on the market before 2 August have additional time to comply with the machine-readable marking requirement.
Deepfakes must be clearly disclosed
Organisations and individuals using AI to produce deepfakes must disclose that the image, audio or video has been artificially generated or manipulated.
The AI Act defines deepfakes as content that resembles real people, objects, places or events and could falsely appear authentic. The rule therefore covers material such as fabricated recordings of politicians, altered videos and artificially generated voices that imitate identifiable people.
Creative, satirical, fictional and artistic works are not fully exempt, but the disclosure may be presented in a way that does not interfere with the normal display or enjoyment of the work.
The rules also cover AI-generated text published to inform the public about matters of public interest. A disclosure is not required when the material has undergone human review or editorial control and a person or organisation accepts editorial responsibility for its publication.
Emotion recognition is also covered
Article 50 extends beyond generative content. Organisations deploying systems for emotion recognition or biometric categorisation must inform the people exposed to them.
This may concern systems that analyse facial expressions, voices or other biometric data to infer emotions or place individuals into categories. The notification requirement operates alongside existing European rules on privacy and personal data protection.
The Commission says the measures are designed to address risks including misinformation, fraud, impersonation and consumer deception. Enforcement will involve national market-surveillance authorities, the European AI Office and, for EU institutions, the European Data Protection Supervisor.
What changes in the Nordic countries
The transparency requirements apply directly in Denmark, Sweden and Finland as EU member states. Denmark’s Agency for Digital Government (Digitaliseringsstyrelsen) has confirmed that the Article 50 obligations apply from 2 August and has published guidance for providers and organisations using AI systems.
The position is different in Norway and Iceland. Both participate in the European Economic Area, but the AI Act has not yet been incorporated into the EEA Agreement. The regulation is still being examined by the EEA EFTA states, meaning its internal application requires a separate incorporation process and national implementation.
Companies based outside the EU may still fall under the regulation when they place AI systems on the European market or when their systems produce outputs used within the EU.
A common standard for synthetic information
The new requirements do not prohibit AI-generated material. Instead, they establish a European principle that people should be able to recognise when artificial intelligence is mediating an interaction or altering the information they receive.
The practical effect will depend on how consistently labels and technical markers are implemented across platforms. The rules nevertheless introduce a shared standard for chatbots, deepfakes and synthetic media, with particular relevance for journalism, online fraud and political communication.
For the Nordic region, they also create a temporary regulatory difference between the three EU member states and Norway and Iceland. That difference is expected to remain until the AI Act is formally incorporated into the EEA framework.
