AI disclosure requirements in Belgium: what applies from 2 August 2026

The EU AI Act's transparency obligations are a Union-wide regulation, so the substance is the same in Belgium as everywhere in the EU: from 2 August 2026, AI systems that interact with people must disclose they're AI, synthetic media must carry machine-readable marking, and deepfakes and certain AI-generated public-interest text must be labeled. What differs country by country is who enforces it, in what language your users encounter it, and how quickly complaints turn into questions — which is what this page covers for Belgium.

Enforcement in Belgium

Belgium designated the BIPT (the federal telecoms regulator, Institut belge des services postaux et des télécommunications / Belgisch Instituut voor postdiensten en telecommunicatie) as its main AI Act regulator in January 2025, but the full national competent-authority structure was still being completed at the time of review — Belgium missed the Act's 2 August 2025 designation deadline and work on the complete supervisory map continues. Article 50 applies from 2 August 2026 regardless, with the EU penalty ceiling of €15M or 3% of worldwide turnover, and Belgium's data protection authority (APD/GBA) remains competent wherever your AI feature processes personal data.

The four duties, as they show up on your site

• Article 50(1) — chatbot disclosure. Any AI system interacting with visitors must make that clear no later than the first interaction, in a clear and distinguishable way. A line in your terms doesn't satisfy it.
• Article 50(2) — machine-readable marking. AI-generated audio, images, video, and text must be marked in a machine-readable, detectable format. For generative systems already on the market before 2 Aug 2026, this specific duty has a transition until 2 Dec 2026 — the rest does not.
• Article 50(3) — emotion recognition & biometric categorisation. People exposed must be informed the system is operating.
• Article 50(4) — deepfake and AI-text labels. Visible disclosure for deepfakes, and for AI-generated or manipulated text published to inform the public on matters of public interest.

Belgium-specific practicalities

• BIPT is a technical regulator with market-surveillance experience from telecoms — machine-readable marking (50(2)) is the kind of check it knows how to run
• Belgium's incomplete designation map does not delay the duties: the regulation binds directly, and cross-border complaints can be coordinated via the Commission's AI Office in the meantime
• Belgium is multilingual by law in consumer matters — plan the disclosure for the language of the storefront (Dutch, French, and German where relevant), not just English
• The APD/GBA complaint route exists today for anything personal-data-adjacent; an undisclosed chatbot that profiles users can arrive as a GDPR complaint long before an AI Act one

Disclosure language should match your user-facing language — in Belgium that often means two: "Je chat met een AI-assistent" for Dutch-language pages and "Vous discutez avec un assistant IA" for French-language ones, each at first interaction.

Common questions