Labelling AI content: what applies from 2 August 2026 under Art. 50 of the AI Act

On 2 August 2026, the transparency obligations under Art. 50 of the AI Act, Regulation (EU) 2024/1689, become applicable. They are aimed at two groups: providers, who develop AI systems and place them on the market, and deployers, meaning companies that use such systems in a professional capacity. If you create texts or images with an AI tool at work, you are generally a deployer.

Obligations of providers

• Chatbots and other systems that interact directly with people must be designed so that users recognise that they are dealing with an AI, unless this is already obvious (Art. 50 Abs. 1).
• Systems that generate synthetic audio, image, video or text content must mark their output in a machine-readable format as artificially generated (Art. 50 Abs. 2).

For generative systems that were already placed on the market before 2 August 2026, the Digital Omnibus Regulation (EU) 2026/1744 has introduced a transitional period until 2 December 2026 (Art. 111 Abs. 4).

Obligations of deployers

Anyone who uses AI to create or distribute a deepfake must disclose that the content has been artificially generated or manipulated (Art. 50 Abs. 4). Deepfakes are image, audio or video content that resembles real people, objects, places or events and would falsely appear to be authentic. For evidently artistic, satirical or fictional works, a notice that does not impair the display of the work is sufficient.

For texts, the obligation applies if they are published to inform the public on matters of public interest. It does not apply if a human has reviewed the text editorially and a person or organisation holds editorial responsibility. An article on the company website drafted with AI, which an employee revises and approves, therefore usually does not need to be labelled. An AI photo that deceptively shows a real person or a real building site, on the other hand, does.

The notice must be clearly recognisable at the latest at the time of the first interaction or exposure and must meet accessibility requirements (Art. 50 Abs. 5).

The code of practice of 10 June 2026

On 10 June 2026, the European Commission published a code of practice on the transparency of AI-generated content. Applying it is voluntary. The first section is aimed at providers and deals with marking and detection. The second concerns deployers and gives guidance on the design, placement and presentation of labels. Annex I contains an optional EU symbol in three variants.

In practice, it makes sense to adopt this symbol for images and videos and to decide internally who reviews and approves AI-generated texts before publication.