On July 20, 2026, the European Commission issued guidelines intended to assist providers and deployers of artificial intelligence (“AI”) systems in complying with the transparency requirements of the EU Artificial Intelligence Act (“AI Act”). The guidelines are intended to clarify how certain disclosure and labelling requirements should operate in practice as the AI Act begins to take effect.

The guidelines address which organizations are subject to the AI Act’s transparency rules and explain several key concepts relevant to compliance, including directly interactive AI systems such as chatbots, synthetic content such as text that is fully or partly generated by AI, and deepfakes.

Under the AI Act, providers of AI systems intended to interact directly with individuals must design and develop those systems in such a way that individuals can be informed that they are interacting with an AI system. Providers must also incorporate machine-readable markings into AI-generated or manipulated content so that the content can be identified as such. These technical measures are intended to support the detection of AI-generated material, including where the content is shared or distributed through digital services.

Deployers of AI systems also have transparency obligations. Deployers must inform individuals when they are exposed to any image, audio or video content that has been AI-generated or manipulated, constituting a deep fake, as well as when they encounter AI-generated or manipulated text concerning matters of public interest that has not been subject to human review or editorial control. Deployers using emotion-recognition or biometric-categorization systems likewise must provide notice to affected individuals in the circumstances covered by the AI Act.

According to the guidelines, organizations may be able to demonstrate compliance with these requirements, as applicable, by voluntarily adhering to the European Commission’s Code of Practice on Transparency of AI-generated content. The European Commission also launched the AI Act Service Desk, which is intended to function as an information resource for organizations seeking assistance in understanding and implementing the AI Act’s requirements.

Most of the transparency provisions are effective as of August 2, 2026. However, providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, placed on the market before that date have until December 2, 2026, to comply with applicable marking and detection requirements.

Read the press release here. Read the Guidelines here.