Source: Pinsent Masons (Out-Law)

The European Commission has issued draft guidance clarifying the transparency obligations under Article 50 of the EU AI Act, and the draft guidelines are open to consultation until 3 June 2026. Article 50 creates a general duty for providers to inform people they intend to interact directly with that they are interacting with an AI system, unless that fact is “obvious from the point of view of a natural person who is reasonably well‑informed, observant and circumspect” or certain law‑enforcement exceptions apply. Providers must also ensure synthetic audio, image, video or text outputs are marked in a machine‑readable format and are detectable as artificially generated or manipulated.

Deployers of AI systems that create or manipulate image, audio or video content constituting deepfakes generally must disclose the artificial nature of the content, subject to limited law‑enforcement exceptions, while content that is evidently artistic, creative, satirical or fictional requires only disclosure of the existence of such generated or manipulated content in a way that does not hamper display or enjoyment. Where AI generates or manipulates text intended to inform the public on matters of public interest, deployers must disclose its artificial origin unless exceptions apply, such as where human review or editorial responsibility exists. All transparency information must be clear and distinguishable at the latest at first interaction or exposure and conform to accessibility requirements. Pinsent Masons lawyers Dr. Nils Rauer and Alexander Bibi welcomed the clarifications but warned that concepts like “obviousness” are anchored in abstract consumer‑law standards that may be difficult for smaller companies to operationalise. Bibi emphasised the Commission’s acknowledgment that no single technical solution currently exists to reliably mark and detect AI‑generated outputs across modalities, making compliance challenging in practice. The guidance expands the definition of “deepfake,” potentially bringing much AI‑generated imagery within scope and forcing case‑by‑case judgments for deployers of advertising and creative content. The guidance is non‑binding while final, raising risks of inconsistent implementation and possible litigation, and the lawyers noted that the Court of Justice of the EU may ultimately determine whether the Commission’s reading of Article 50 strikes the right balance. The Commission has also published a code of practice in parallel to support providers and deployers on marking and labelling AI‑generated outputs.