Foundations
Prohibited AI practices (Article 5)
Eight AI practices are banned outright under the Act — in force since February 2025, and carrying the heaviest penalties. A plain-language list.
Updated 3 July 2026 · Reference material, not legal advice
At the top of the risk pyramid sit practices considered an unacceptable risk to people and society. These are prohibited outright by Article 5. They have applied since 2 February 2025, and they carry the Act’s heaviest penalties: up to €35 million or 7% of worldwide annual turnover, whichever is higher.
The eight prohibited practices
- Subliminal, manipulative or deceptive techniques that materially distort behaviour and cause significant harm.
- Exploitation of vulnerabilities — of age, disability, or a specific social or economic situation — to distort behaviour harmfully.
- Social scoring — evaluating or classifying people over time on social behaviour or personal characteristics, leading to unjustified or disproportionate detrimental treatment.
- Predictive policing on individuals — assessing the risk of a person committing an offence based solely on profiling or personality traits. (Systems that support a human assessment grounded in objective, verifiable facts are not caught.)
- Untargeted scraping of facial images from the internet or CCTV to build or expand facial-recognition databases.
- Emotion recognition in the workplace and in education — save for medical or safety reasons.
- Biometric categorisation that infers sensitive attributes (race, political opinions, trade-union membership, religious or philosophical beliefs, sex life or sexual orientation).
- ‘Real-time’ remote biometric identification in public spaces for law enforcement — prohibited in principle, with narrowly drawn exceptions subject to prior authorisation and safeguards.
Why this matters even if you think it does not apply
Two points catch teams out. First, some of these can be triggered inadvertently — an HR analytics feature that infers mood, or a marketing tool that manipulates. Second, the prohibitions are already in force, unlike most high-risk obligations that phase in later. A quick check against Article 5 is worth doing early.
What this means in practice
Euridium’s assessment flags prohibited-practice exposure at the start, so a banned use is caught before it becomes a finding — not after.
Run a guided assessment and get the obligations that apply to your role.
This guide is a plain-language summary for orientation. The authoritative text is Regulation (EU) 2024/1689 as published in the Official Journal of the European Union (12 July 2024). It does not constitute legal advice.