Act

An EU code is being introduced to ensure the transparency of AI-generated content

In view of the requirements coming into force on 2 August, the Commission has published guidelines on labelling GenAI videos, images and text

by Camilla Colombo

11 June 2026

FILE PHOTO: FILE PHOTO: A copy of "The European Union Artificial Intelligence (AI) Act" on display during the AI & Big Data Expo 2025 at the Olympia, in London, Britain, February 5, 2025. REUTERS/Isabel Infantes/File Photo/File Photo REUTERS

3' min read

Translated by AI

Versione italiana

Key points

3' min read

Translated by AI

Versione italiana

A set of best practices to help ensure better compliance with the transparency requirements of the Artificial Intelligence Act, which will come into force on 2 August. Thus, the European Commission gave the go-ahead today, Wednesday 10 June, for the publication of the code of good practice on marking and labelling of AI-generated content. The code is voluntary and sets out concrete measures to assist both providers and operators of generative artificial intelligence systems.

Aiming for transparency

From 2 June 2026, the date on which the rules and obligations introduced by EU Regulation 2024/1689 will come into full effect, the AI Act will require clear labelling for certain particularly sensitive content, such as deepfakes and texts generated by AI, or manipulated by AI, and published on matters of public interest. Users must also be clearly informed when they interact with a interactive AI system, such as a chatbot.

This set of guidelines, based on transparency, has a specific aim: to help people recognise when content has been generated or modified by artificial intelligence, thereby reducing the risk of deception, manipulation, disinformation or misinformation.

The code

The code was drawn up by six independent experts, with input from over 180 stakeholders, including providers and operators of interactive and generative AI systems and models; associations representing operators; small and medium-sized enterprises; the academic community; the public sector; and civil society organisations.

The Code consists of two sections: suppliers and distributors. The first section focuses on the obligations that suppliers of generative artificial intelligence systems must comply with. It sets out how they must ensure that audio, images, video or text generated or manipulated by AI are labelled in a machine-readable manner and can be understood, identified and recognised as having been artificially generated or manipulated by all users. The second section is dedicated to the obligations of distributors of generative AI systems. It explains how they must clearly label deepfakes and AI-generated or AI-manipulated texts intended to inform the public on matters of general interest, particularly where there has been no human review or editorial control.

These transparency rules complement those of the AI Act on general-purpose artificial intelligence models and on high-risk systems. The European Commission’s objective, in fact, is the responsible development and use of this innovation within the European context, even at the risk of falling behind the efforts of the United States and China. ‘European citizens have the right to know whether what they see, hear or read has been created or modified by artificial intelligence, especially when such content can influence and shape the public debate. Transparency is how we protect trust,” comments Henna Virkkunen, Executive Vice-President for Technological Sovereignty, Security and Democracy. “We call on providers and operators to sign up to the code and show leadership in the development and implementation of responsible and trustworthy AI.”

Starting 2 June

The code is now open for signatures from all interested parties and will be further supplemented by Commission guidelines designed to clarify the scope of the legal obligations and address aspects not covered by the code. The guidelines will, for the most part, be practical in nature, designed to help suppliers and distributors meet transparency requirements. Those who sign it will thus be able to demonstrate compliance with the obligations of the AI Act, which will come into force on 2 August.