EU AI Act · transparency deadline

Article 50: what chatbots & generative AI must disclose from 2 August 2026

The Digital Omnibus moved the high-risk deadlines to 2027–28 — but not this one. From 2 August 2026, chatbots must say they are AI, synthetic content must carry machine-readable marking, deepfakes and AI-written public-interest text must be labelled, and emotion-recognition deployments must be disclosed. Fines reach €15M or 3% of worldwide turnover.

⏰ 20 days until Article 50 applies on 2026-08-02.

Art. 50(1) — Tell people they are talking to an AI

Design the system so the people interacting with it are informed that they are interacting with an AI system, in a clear and distinguishable way, at the latest at the first interaction.

Provider duty (you build / place it on the market) · Exemption: Not required where it is obvious to a reasonably well-informed, observant and circumspect person from the context, or for AI systems legally authorised to detect, prevent or investigate criminal offences.

Art. 50(2) — Mark synthetic content in a machine-readable way

Ensure outputs of the system (audio, image, video or text) are marked in a machine-readable format and detectable as artificially generated or manipulated. Solutions (e.g. watermarks, metadata such as C2PA, cryptographic provenance, fingerprinting) must be effective, interoperable, robust and reliable as far as technically feasible.

Provider duty (you build / place it on the market) · Exemption: Not required where the system performs an assistive function for standard editing or does not substantially alter the input data, or for systems legally authorised for criminal-offence purposes.

Art. 50(3) — Disclose emotion recognition / biometric categorisation

Inform the natural persons exposed to the system that an emotion-recognition or biometric-categorisation system is operating, and process their personal data in line with GDPR.

Deployer duty (you operate it) · Exemption: Limited law-enforcement carve-out where permitted by law with appropriate safeguards. Note: emotion inference at the workplace or in education is prohibited outright under Art. 5 — this paragraph cannot legalise those uses.

Art. 50(4) ¶1 — Label deepfakes

Disclose that image, audio or video content constituting a deepfake has been artificially generated or manipulated.

Deployer duty (you operate it) · Exemption: For evidently artistic, creative, satirical or fictional work, the duty is limited to a disclosure that does not hamper the display or enjoyment of the work. Law-enforcement carve-out applies.

Art. 50(4) ¶2 — Label AI text published to inform the public

Disclose that text published with the purpose of informing the public on matters of public interest has been artificially generated or manipulated.

Deployer duty (you operate it) · Exemption: Not required where the AI-generated content has undergone human review or editorial control and a natural or legal person holds editorial responsibility for its publication.

Check your duties free — get disclosure text & a compliance note → Free AI policy templates

The wizard runs entirely in your browser — no sign-up, nothing uploaded. It produces your exact Article 50 duty list, ready-to-paste disclosure text, an implementation checklist and a downloadable compliance note. Educational aid, not legal advice.