European Commission consults on Code of Practice on AI transparency obligations under the AI Act

The question

How do the transparency obligations under the EU AI Act affect businesses using AI-generated content? What impact will this have on the advertising and marketing industry?

The key takeaway

The EU Commission is consulting on a new Code of Practice for Businesses who use or plan to use AI-generated content should ensure that they comply with the transparency obligations under the AI Act. A key aspect will be labelling AI-generated content in a clear and distinguishable manner at the point of users’ first exposure. The Code of Practice, expected by mid-2026, will provide practical guidance in advance of Article 50’s application on 2 August 2026.

The background

The AI Act entered into force on 1 August 2024, establishing an EU-wide framework to regulate AI systems based on the level of risk they pose. Article 50 introduces transparency rules for deployers and providers, requiring companies to disclose image, audio, or video content that has been “generated or manipulated” by an AI system, and to inform users when they are interacting with AI systems. The Act has extraterritorial reach and may apply to entities outside the EU. See our Summer 2024 Snapshots for more.

Deepfakes under the AI Act

Under Article 50 of the AI Act, a deepfake is defined as “AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful”. This definition is broad and likely catches promotional materials, such as ads, which have been produced or significantly altered using AI.

Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake must disclose that the content has been artificially generated or manipulated through appropriate labelling. This disclosure must be provided in a clear and distinguishable manner at the latest at the time of the first interaction or exposure. The disclosure must also comply with applicable accessibility requirements.

There are exceptions: (i) where the AI-generated content is for creative, satirical or artistic purposes; or (ii) in journalistic circumstances where the AI-generated content is to inform the public on matters of public interest and is under editorial control.

The development

In September 2025, the European Commission launched a consultation inviting stakeholders to share their views on Article 50 with a view to producing a Code of Practice. This is pursuant to Article 50(7) that allows the Commission to produce codes of practice to facilitate the effective implementation of the labelling obligations under the Act. If it deems the Code is not adequate, the Commission may go further and adopt common rules for the implementation of these obligations. The Commission expects that, following the consultation, the drafting process of the Code of Practice will be complete by June 2026.

Impact on advertising and marketing: EU vs UK

The AI Act’s prescriptive transparency obligations differ from the UK Advertising Standards Authority (ASA)’s more principles-based approach. The ASA focuses on whether the absence of disclosure would mislead the audience. Its recent guidance advises businesses using AI-generated ads to consider:

• whether the audience is likely to be misled if the use of AI is not disclosed, and
• if disclosure is required, whether it clarifies or contradicts the ad’s message.

This represents a less rigid approach than the EU’s explicit labelling requirements under the AI Act.

Why is this important?

Although the practical guidance in the Code of Practice will be voluntary, it is likely to give clarity on the steps businesses should take to comply with the transparency rules. Advertisers will need to consider how to appropriately label and disclose that AI has been used to generate an ad without detracting from the content, and how such disclosures can feasibly be made across different types of media. The AI Act provides for significant administrative fines for severe breaches of the transparency obligations: up to the higher of €15 million or 3% of global annual turnover. This underscores the importance of proactive compliance and ongoing monitoring of regulatory developments.

Any practical tips?

The Code of Practice will cover all AI systems generating/manipulating content, interacting with people, or classifying biometric/emotional information. From a business perspective, one of the immediate impacts of Article 50 will be on the use of AI in advertising and marketing, so the Code will make essential reading for those looking to use AI generated content in EU-facing campaigns. While Article 50 does not land with full effect until August 2026, businesses should already be assessing to what extent they are using AI generated content in their advertising and, if so, how they can adequately disclose the use of AI. Creation of AI governance policies, if not already in place, will assist in a consistent approach to labelling content across multiple types of media.

Autumn 2025