Back to home

EU AI Act: Transparency obligations for AI-generated content apply from 2 August 2026

Category: IP/IT/Data

AI Laws and Regulations Concept. Wooden cube with AI and law icons on the Computer motherboard. regulation, standard and business policy. Laws to ensure that AI is safe, transparent, and traceable.

As of 2 August 2026, deployers of certain AI systems will be required to disclose that specific content has been artificially generated or manipulated.

These obligations are set out in Article 50(4) of the EU Regulation 2024/1689 on artificial intelligence (“AI Act”).

Article 50 also contains separate obligations for providers of AI systems.

Key takeaways

  • The disclosure obligations apply from 2 August 2026.

  • They do not apply to every use of generative AI. For deployers, disclosure is required where an AI system is used to generate or manipulate (i) image, audio or video content constituting a deep fake; or (ii) )text published for the purpose of informing the public on matters of public interest.

  • AI-generated public-interest text does not require disclosure where it has undergone substantive human review or editorial control and a natural or legal person assumes editorial responsibility for its publication.

  • The disclosure must be clear and distinguishable and provided, at the latest, when a person is first exposed to the content.

  • The EU transparency icons may be used voluntarily, but their use does not, in itself, establish compliance.

A.   When is an organisation a deployer?

Under the AI Act, a deployer is a natural or legal person, public authority, agency or other body that uses an AI system under its authority, except where the system is used in the course of a personal, non-professional activity.

For example, a company whose marketing or communications team operates a generative AI tool under the company’s authority will, in principle, be the deployer. Where an external agency uses its own AI systems independently and supplies only the finished content, the respective roles of the company and the agency will need to be assessed in light of the specific arrangements between them.

B.   What type of content is concerned?

Once an organisation qualifies as a deployer, it must determine whether the content concerned falls within either of the two categories addressed below.

1)    Deep fakes

Deployers must disclose that image, audio or video content has been artificially generated or manipulated where it constitutes a deep fake. 

Under the AI Act, a deep fake is AI-generated or manipulated content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful. This is for instance the case for an AI-generated video featuring a realistic synthetic avatar of a company CEO congratulating employees with their work and the corporate results.

Where a deep fake forms part of an evidently artistic, creative, satirical, fictional or analogous work or programme, the disclosure obligation is more limited. In such cases, the existence of the AI-generated or manipulated content must be disclosed in an appropriate manner that does not hamper the display or enjoyment of the work.

2)    Text published on matters of public interest

Deployers must also disclose that text has been artificially generated or manipulated where it is:

  • published;

  • intended to inform the public; and

  • concerned with a matter of public interest.

Matters of public interest may include politics and democratic processes, public administration, justice and law enforcement, fundamental rights, public security, public health, environmental protection, consumer safety, and economic, financial, scientific or cultural developments that may be relevant to public debate.

An AI-generated article addressing the safety of a consumer product, the health effects of an ingredient or a significant environmental development may therefore fall within the disclosure obligation.

However, disclosure is not required where the text has undergone human review or editorial control, and a natural or legal person assumes editorial responsibility for its publication (cumulative conditions). Superficial or purely formal checks, such as correcting spelling, grammar or formatting, are not sufficient.

C.   How should the disclosure be made?

The disclosure must be provided in a clear and distinguishable manner, at the latest when a person is first exposed to the content. It must also comply with applicable accessibility requirements.

The information must be understandable and perceptible without requiring the audience to use a particular technical tool or take additional steps. Depending on the medium, this may require e.g.:

  • a visible label placed directly on an image or video;

  • an audible disclosure accompanying audio content; or

  • a prominent notice presented with a published text.

The Commission has also developed three voluntary EU transparency icons to indicate when a content has been fully or partially made with AI.

The icons are available in different formats and can be downloaded in SVG and PNG formats.

Non-compliance with the transparency obligations laid down in Article 50 may result in administrative fines of up to EUR 15 million or, where the infringer is an undertaking, up to 3% of its total worldwide annual turnover for the preceding financial year, whichever is higher (Article 99(4)(g) of the AI Act). Please note that, to date, no Belgian authority has been formally and publicly designated to enforce these obligations and impose the corresponding fines.

D.   What about providers?

A provider is a person or an organisation that develops an AI system, or has one developed, and places it on the market or puts it into service under its own name or trademark.

Article 50 also imposes separate obligations on providers of AI systems, such as the obligation to ensure that outputs generated or manipulated by AI systems are marked in a machine-readable format and are detectable as artificially generated or manipulated.

What should businesses do now?

Businesses acting as deployers and using AI to generate or manipulate public-facing content should review their practices by 2 August 2026.

In particular, they should:

  • identify the AI systems used in their content-production processes and determine under whose authority they are operated;

  • identify deep fakes and relevant public-interest publications;

  • establish appropriate editorial-review and labelling procedures; and

  • clearly allocate responsibilities where agencies, contractors or other external content providers are involved.

Our team is available to assist businesses in determining how Article 50 applies to their activities and in implementing measures.

Subscribe to Our newsletter

Back to overview