Artificial Intelligence: the new transparency requirements come into force on 2 August

From 2 August 2026, the transparency obligations set out in Article 50 of Regulation (EU) 2024/1689 — the AI Act — will apply. These rules impose specific obligations on providers and professional users of artificial intelligence systems operating in the European market.

To support the practical implementation of these obligations, the European Commission published, in July 2026, the Guidelines on Transparency Obligations and the Code of Practice on Transparency of AI-Generated Content — two instruments that clarify what is actually required of organisations.

It is now clearer that anyone who uses or makes AI systems available must ensure transparency towards users and the general public.

We explain below what is changing.

What is changing? 

Who is covered? 

• Providers — organisations that develop or place AI systems on the market or put them into service within the European Union.
• Professional users (deployers) — organisations that use AI systems under their authority in the course of their professional activities (excluding personal use of a non-professional nature).

When is it mandatory to inform the user that they are interacting with AI? 

• Providers of AI systems designed to interact directly with people — such as chatbots, virtual assistants or avatars — must ensure that the user is informed that they are interacting with an AI system.
• This obligation may only be waived where this is obvious, taking into account the circumstances and context of use.

When must AI-generated content be identified? 

• Providers of AI systems that generate synthetic content — including text, images, audio or video — must ensure that such content is labelled in a machine-readable format in an effective, interoperable, robust and reliable manner, to the extent that this is technically feasible.
• Professional users of systems that generate text published with the aim of informing the public on matters of public interest must disclose that the content has been generated or manipulated by AI — unless the content has been subject to human review or editorial control and there is a natural or legal person with editorial responsibility for the publication.

What are the rules on deepfakes? 

• Anyone using AI systems to generate or manipulate images, audio or video that constitute deepfakes is required to disclose that the content has been artificially generated or manipulated.
• The AI Act provides for limited exceptions, notably for artistic or fictional contexts.

Additional obligations for professional users (deployers) 

• Anyone using emotion recognition or biometric categorisation systems must inform individuals who are exposed to the operation of such systems.

Key clarifications from the Guidelines and the Code of Practice 

• The Guidelines published by the European Commission on 20 July 2026 define the scope of the obligations and provide practical guidance for providers, professional users and competent authorities.
• The Code of Practice is a voluntary instrument that supports compliance with the obligations regarding the flagging and labelling of AI-generated content. Those adhering to the Code of Practice may use it to demonstrate compliance with the obligations under Article 50, whilst those opting for other measures will have to demonstrate that they are equally appropriate.
• Annex I of the Code of Practice includes a set of threeEU icons which professional users (deployers) may freely use to label AI-generated or AI-manipulated content, in accordance with Article 50(4) of the AI Act.
• These icons form an integral part of Section 2 of the Code of Practice and have been validated through empirical user testing in several Member States; they are intended to help people recognise, clearly and distinctively, that content has been artificially generated or manipulated by AI.
• Each icon is available in four variations (black, white, black with 50% transparency and white with 50% transparency), in SVG and PNG formats, and is free to use.
• The three icons set out in Annex I to the Code of Practice are as follows:

• Section 1 of the Code of Practice sets out the commitments of providers regarding the labelling and detection of generated or manipulated content. Section 2 sets out the commitments of professional users regarding the labelling of deepfakes and AI-generated text of public relevance.
• As regards the process for adhering to the Code of Practice, this is voluntary, and providers and professional users of generative AI systems who wish to demonstrate compliance with the labelling and tagging obligations set out in Article 50 of the AI Act may adhere to it.
• Adherence is effected by completing an adherence form, which must be submitted by email and signed by a person with sufficient authority to bind the provider to the Code of Practice.
• The legal effects of adherence are as follows: adherents may rely on the measures set out in the Code of Practice to demonstrate compliance with the obligations under Article 50 of the AI Act, thereby benefiting from greater predictability, legal certainty and trust amongst Member States. Those who do not adhere to the Code of Practice may fulfil their obligations by other means, but will have to demonstrate that those means are equally appropriate.

Consequences of non-compliance 

The AI Act provides for fines of up to €15 million or up to 3 per cent of the company’s global annual turnover (whichever is higher) for breaches of transparency obligations. For SMEs and start-ups, the regulation provides for adjusted thresholds, with the lower amount (rather than the higher) applying.

Recommended actions 

For organisations that use or make available artificial intelligence systems, we recommend the following actions before 2 August 2026:

1. Identify the AI systems used within the organisation, including chatbots, content generation tools, emotion recognition systems and biometric categorisation systems.
2. Check whether any transparency obligations apply, based on the type of system and the capacity in which the organisation operates (as a provider or a professional user).
3. Review internal policies and procedures, ensuring that information and disclosure mechanisms are in place and adequate.
4. Assess adherence to the Code of Practice as a means of demonstrating compliance with the obligations regarding the flagging and labelling of AI-generated content.
5. Adapt processes before the rules come into force, ensuring that the organisation is compliant from 2 August 2026.

The Digital Contracting & Compliance team at Antas da Cunha Ecija closely monitors regulatory developments in the field of artificial intelligence and is available to support organisations in assessing the impact of these new obligations and in implementing the necessary measures to ensure compliance.