Commission Guidelines on prohibited artificial intelligence practices established by Regulation (EU) 2024/1689 (AI Act)
This document provides non-binding guidelines from the European Commission on the practical implementation of prohibited artificial intelligence practices established by Regulation (EU) 2024/1689 (the AI Act). It covers the rationale, scope, exceptions, and enforcement mechanisms for prohibitions related to harmful manipulation, deception, social scoring, criminal risk assessment, facial image scraping, emotion recognition, biometric categorization, and real-time remote biometric identification. The guidance aims to ensure consistent, effective, and uniform application of Article 5 of the AI Act while balancing fundamental rights protection and innovation.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
Competent authorities, AI providers, and AI deployers seeking practical guidance on complying with and enforcing the prohibited practices provisions of the AI Act.
What the document contains
The document addresses the practical interpretation, material and personal scope, exclusions, and exemptions associated with the prohibited AI practices listed under Article 5 of Regulation (EU) 2024/1689. It details specific components such as subliminal techniques, deceptive practices, exploitation of vulnerabilities, social scoring, crime prediction, untargeted facial scraping, emotion recognition in workplaces and schools, sensitive biometric categorization, and real-time remote biometric identification for law enforcement.
It does not provide binding legal rulings, as authoritative interpretations of the AI Act remain the exclusive competence of the Court of Justice of the European Union. It also does not cover detailed operational checklists for high-risk AI system conformity assessments outside the scope of Article 5.
- The guidelines interpret the prohibited AI practices under Article 5 of the AI Act to provide legal clarity and ensure uniform application across EU Member States.
- Article 5 prohibitions cover specific practices considered unacceptably harmful to fundamental rights and EU values, including subliminal manipulation, social scoring, and real-time remote biometric identification in public spaces.
- The guidelines clarify both provider and deployer responsibilities across the AI lifecycle, emphasizing the need for proportionality, risk assessment, and human oversight.
- Strict, narrowly defined exceptions apply to certain prohibitions—such as law enforcement use of real-time RBI—subject to prior authorization and robust fundamental rights safeguards.
- The AI Act applies horizontally without prejudice to other EU legislation, such as data protection (GDPR, LED) and consumer protection laws.
1. Background and objectives 2. Overview of prohibited AI practices 3. Article 5(1)(a) and (b) AI Act – harmful manipulation, deception and exploitation 4. Article 5(1)(c) AI Act - social scoring 5. Article 5(1)(d) AI Act – individual risk assessment and prediction of criminal offences 6. Article 5(1)(e) AI Act - untargeted scraping of facial images 7. Article 5(1)(f) AI Act emotion recognition 8. Article 5(1)(g) AI Act: Biometric categorisation for certain 'sensitive' characteristics 9. Article 5(1)(h) AI Act - Real-time Remote Biometric Identification (RBI) Systems for Law Enforcement Purposes 10. Safeguards and Conditions for the exceptions (Article 5(2)-(7) AI Act) 11. Entry into application 12. Review and update of the Commission guidelines
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Human Dignity
5/5Respect for the inherent worth of persons
The prohibitions aim to protect the right to human dignity (Article 1 of the Charter), which also constitutes the basis of all fundamental rights and includes individual autonomy as an essential aspect.
3.1
screening score · 2026-10-24
Privacy
4/5Protection of personal data and private life
The untargeted scraping of facial images from the internet and from CCTV footage seriously interferes with individuals’ rights to privacy and data protection and deny those individuals the right to remain anonymous.
6.1
screening score · 2026-10-24
Transparency
4/5Openness about how AI systems work and are used
provide transparency in how the AI system operates, clear disclosures about its capabilities and limitations, and relevant information to ensure informed decisions
3.2.3
screening score · 2026-10-24