Commission publishes guidelines on prohibited AI practices under the AI Act

On 4 February 2025 the European Commission published guidelines on prohibited AI practices under Regulation (EU) 2024/1689. They interpret the prohibitions that have applied since 2 February 2025 — including harmful manipulation, social scoring and real-time remote biometric identification — to support consistent application across the EU. The guidelines are non-binding; authoritative interpretation rests with the Court of Justice.

The prohibitions are the AI Act's only outright bans and carry its highest penalty tier. These guidelines are the Commission's own reading of where the lines sit — the reference point any assessment will be measured against.

Providers and deployers of AI systems that could touch a prohibited practice: behavioural targeting, scoring, biometric or emotion-recognition use cases in particular.

Screen your AI portfolio against the prohibited-practice categories as interpreted in the guidelines, and document the conclusion per system — including the borderline calls.