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

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

JURI-AD-719827

Committee on Legal Affairs · Rapporteur: Axel Voss

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.

The Committee on Legal Affairs proposes amendments to the proposed Artificial Intelligence Act. It adds general principles for all AI systems, AI literacy duties, and rights to explanation, complaint and judicial remedy for affected persons. It narrows the scope for military AI, research, business-to-business use and open source systems, and adds rules for general purpose AI and open source AI. It strengthens transparency, human oversight and user duties for high-risk AI systems, and requires providers to report serious incidents within 72 hours. It turns the European Artificial Intelligence Board into an independent body with a secretariat and gives it tasks including binding decisions in cross-border conflicts. It requires each member state to designate one national supervisory authority and sets up a consistency mechanism for cross-border cases.

Committee position. The Committee on Legal Affairs calls on the lead committees to take into account its amendments to the proposed Artificial Intelligence Act. The amendments add general principles, AI literacy, rights for affected persons, and changes to scope, transparency, oversight, governance and remedies.

Key points

  1. Adds a new Article 4a with general principles for all AI systems: human agency and oversight, technical robustness and safety, privacy and data governance, and transparency.
  2. Adds a new Article 4b requiring the Union, member states, providers and users to promote AI literacy through education, training and reskilling.
  3. Deletes the exclusion of AI systems developed or used exclusively for military purposes from the regulation's scope.
  4. Exempts research, testing and development before market placement or service, and AI systems used in a strictly business-to-business environment that pose no risk.
  5. Exempts open source AI systems until they are put into service or made available for payment, and defines general purpose AI systems and open source AI systems.
  6. Requires high-risk AI systems to be transparent so providers and users can understand their functioning, and to come with intelligible instructions for use.
  7. Requires human oversight of high-risk AI systems by persons with sufficient AI literacy, support and authority, and a 'stop' button that halts the system safely.
  8. Requires users of high-risk AI systems to follow instructions, assign oversight to persons with AI literacy, ensure relevant input data, monitor operation and keep logs for at least six months.
  9. Requires providers to report serious incidents or malfunctioning without undue delay and no later than 72 hours after becoming aware.
  10. Establishes the European Artificial Intelligence Board as an independent body with its own legal personality, secretariat, strong mandate and sufficient resources.
  11. Creates a consistency mechanism for cross-border cases, with a lead national supervisory authority and binding Board decisions within three months of notification of a serious disagreement.
  12. Grants affected persons a right to an explanation for AI decisions, a right to lodge complaints with supervisory authorities, and a right to effective judicial remedy.

Who is affected

  • Providers and users of AI systems, who must follow new general principles, AI literacy duties and transparency and oversight rules.
  • Affected persons, who gain rights to an explanation, to complain to a supervisory authority and to an effective judicial remedy.
  • Small and medium-sized enterprises and start-ups, for which the Board must provide specific guidance and which benefit from innovation support measures.
  • Member states, which must designate one national supervisory authority with adequate financial and human resources.
  • Open source AI providers, whose systems are exempt until put into service or made available for payment.

Figures and deadlines

  • Open Source software contributes between €65bn -€95bn to the European Union's GDP.
  • Serious incidents or malfunctioning must be reported no later than 72 hours after the provider becomes aware.
  • Users must keep logs for a period of at least six months, unless provided otherwise in applicable Union or national law.
  • The Board shall issue a binding decision within three months of notification of a serious disagreement between national supervisory authorities.
  • The Commission should consider whether an EU Agency is necessary within three years after the date of application of the regulation.
  • The Commission shall evaluate the effectiveness of the Board and assess whether an EU Agency would be best placed within three years after the date of application.