AI disclosure requirements in Romania: 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 Romania 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 Romania.

Enforcement in Romania

Romania designated its national AI Act authorities by government memorandum on 12 March 2026 — about seven months after the EU's original deadline. ANCOM (the National Authority for Administration and Regulation in Communications) was named both the market surveillance authority and the single national contact point. ANSPDCP (the National Supervisory Authority for Personal Data Processing) holds supervisory authority in sensitive areas — biometrics, law enforcement, migration, asylum, border control, justice, and democratic processes — and remains the competent authority wherever an AI system's personal-data processing is the core issue. The financial sector gets its own lane: the Financial Supervisory Authority (ASF) and the National Bank of Romania (BNR) are the proposed supervisors for high-risk AI used by the institutions they already regulate. Article 50 applies from 2 August 2026 with the EU's penalty ceiling of €15 million or 3% of worldwide turnover.

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.

Romania-specific practicalities

• ANCOM as both market surveillance authority and single point of contact gives Romania one clear front door for chatbot-disclosure and AI-labelling complaints outside the carved-out sensitive-data and financial lanes
• If your AI feature touches biometrics, law-enforcement-adjacent data, or other sensitive categories, ANSPDCP — not ANCOM — is the authority to watch
• Romania's seven-month delay past the EU designation deadline doesn't delay Article 50 itself; the duties were live on schedule regardless of when the national memorandum landed
• Keep the Romanian-language first-interaction screenshot in your evidence file — a notice your Romanian users can't read is a weak notice

Disclosure language should match your user-facing language. For a site serving Romania in Romanian, a first-message line such as "Discuți cu un asistent AI" (you're chatting with an AI assistant) is the pattern to aim for — at first interaction, not in termeni și condiții.

Common questions