Skip to content
REPORT (COM(2021)0206 – C90146/2021 – 2021/0106(COD))

On the proposal for a regulation of the European Parliament and of the Council on laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts

A-9-2023-0188

Committee on the Internal Market and Consumer Protection Committee on Civil Liberties, Justice and Home Affairs · Rapporteur: Brando Benifei, Ioan-Dragoş Tudorache

Official HTML ↗ PDF DOCX
✦ In shortFull textProcedure & committee votes
✦ In short · AI summary of this text, generated 26 Sept 2026

Written by a language model from the full text only; every figure comes from the text. The paragraph links open the paragraph a statement rests on — check the text itself before relying on it.

Parliament's amended version of the Commission's proposed Artificial Intelligence Act sets a risk-based framework: it bans certain AI practices, lays down requirements and obligations for high-risk AI systems, and imposes transparency duties on other systems. It adds prohibitions on social scoring, biometric categorisation, predictive policing, untargeted facial scraping and emotion detection in law enforcement, border management, workplaces and schools, and bans 'real-time' remote biometric identification in publicly accessible spaces. It sets obligations for providers, deployers, importers and distributors, including risk management, data governance, technical documentation, logging, human oversight, accuracy and cybersecurity, plus a fundamental rights impact assessment for deployers. It creates an EU Artificial Intelligence Office, national supervisory authorities, an EU database, regulatory sandboxes and rules on remedies, complaints and fines. It adds rules for foundation models, measures for SMEs and start-ups, and protection for workers, research and free and open-source AI components.

Committee position. The committees propose adopting Parliament's first-reading position with numerous amendments to the Commission proposal, adding prohibitions, high-risk classifications, operator obligations, governance structures and remedies.

Key points

  1. The regulation aims to promote human-centric and trustworthy AI and ensure a high level of protection of health, safety, fundamental rights, democracy, rule of law and the environment while supporting innovation.
  2. It prohibits AI practices using subliminal or manipulative techniques, exploiting vulnerabilities, biometric categorisation by sensitive attributes, and social scoring.
  3. It bans 'real-time' remote biometric identification in publicly accessible spaces, predictive policing, untargeted facial scraping and emotion detection in law enforcement, border management, workplaces and education.
  4. High-risk AI systems must meet requirements on risk management, data governance, technical documentation, logging, transparency, human oversight, accuracy, robustness and cybersecurity.
  5. Providers must operate a quality management system, carry out conformity assessment, draw up an EU declaration of conformity, affix CE marking and set up post-market monitoring.
  6. Deployers must use systems per instructions, monitor operation, keep logs, inform affected persons, and conduct a fundamental rights impact assessment before putting certain high-risk systems into use.
  7. Providers of foundation models must assess and mitigate risks, apply data governance, meet performance and cybersecurity requirements, and provide technical documentation to downstream providers.
  8. Transparency obligations require informing people they interact with AI, disclosing emotion recognition and biometric categorisation, and labelling deep fakes.
  9. An independent European Artificial Intelligence Office is established, with a management board, secretariat and advisory forum, to support consistent application.
  10. Member States must designate national supervisory authorities with adequate resources, and these act as market surveillance authorities.
  11. Member States must establish at least one AI regulatory sandbox, with priority access for SMEs and start-ups, and the Commission may adopt common specifications.
  12. Affected persons get rights to complain, to an effective judicial remedy and to an explanation of individual decisions, and fines are set for infringements.

Who is affected

  • Providers, deployers, importers and distributors of AI systems in the EU face new obligations and liability.
  • SMEs and start-ups get priority sandbox access, reduced conformity assessment fees and protection from unfair contract terms.
  • Workers and their representatives must be consulted before high-risk AI is used at the workplace.
  • Law enforcement, migration and border authorities face bans and restrictions on biometric and emotion-detection AI.
  • Affected persons can complain, seek judicial remedy and request explanations for AI decisions.

Figures and deadlines

  • Fines up to 40 000 000 EUR or 7% of worldwide annual turnover for prohibited practices.
  • Fines up to EUR 20 000 000 or 4% of turnover for non-compliance with data and transparency requirements.
  • Fines up to EUR 10 000 000 or 2% of turnover for other non-compliance.
  • Serious incidents must be reported not later than 72 hours after awareness.
  • Logs must be kept for at least 6 months.
  • National supervisory authorities must be designated within three months after entry into force.
  • Joint investigations may be triggered where an infringement affects at least 45 million individuals in more than one Member State.
  • Conformity certificates shall not exceed four years.

Legal basis: Article 294(2) and Articles 16 and 114 of the Treaty on the Functioning of the European Union.