Skip to content
DRAFT 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

ITRE-PA-719801

Committee on Industry, Research and Energy · Rapporteur: Eva Maydell

Official HTML ↗ PDF DOCX
✦ In shortFull text
✦ 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 Industry, Research and Energy's draft opinion responds to the Commission's proposed Artificial Intelligence Act. It proposes amendments to support innovation, especially for start-ups and SMEs, through stronger regulatory sandboxes and an EU AI Regulatory Sandboxing Programme. It calls for an internationally recognised definition of AI system aligned with the OECD, and high but realistic standards for accuracy, robustness, cybersecurity and data, supported by a common European benchmarking authority. It seeks to encourage AI uptake by fostering social trust and clarifying value chain responsibility, including rules for general purpose AI systems and clear disclosure of deep fakes. It aims to future-proof the Act by linking it to the green transition, monitoring market trends and foresight, and involving the High Level Expert Group on AI.

Committee position. The rapporteur proposes amendments to the Commission's AI Act to support innovation, especially for start-ups and SMEs, through stronger regulatory sandboxes, an internationally recognised definition of AI system, realistic standards, and clearer value chain responsibility.

Key points

  1. The draft opinion supports the Commission's horizontal risk-based approach to AI and proposes four objectives: innovation support, a clear definition and realistic standards, uptake and social trust, and future-proofing.
  2. It proposes strengthening regulatory sandboxes with a European dimension and an EU AI Regulatory Sandboxing Programme, with modalities set out in a new Annex, and priority access for SMEs and start-ups.
  3. It calls for an internationally recognised definition of AI system aligned with the OECD, and for continued dialogue between the EU and the OECD as understanding of AI systems evolves.
  4. It calls for high but realistic standards for accuracy, robustness, cybersecurity and data, and for a common European authority on benchmarking bringing together national metrology and benchmarking authorities.
  5. It seeks to foster social trust through a collaborative relationship between developers and users, stronger Codes of Conduct, and compliance-by-design measures.
  6. It proposes rules for general purpose AI systems: they are not subject to the Regulation on their own, but a person placing one on the market for an intended purpose becomes the provider.
  7. It proposes that users of deep fakes must clearly and explicitly disclose that content has been artificially generated or manipulated.
  8. It proposes that the European Artificial Intelligence Board monitor digital skills, carry out annual horizon scanning and foresight, and publish annual recommendations to the Commission.
  9. It proposes that the Commission and the Board include provisions for general purpose AI systems in Codes of Conduct, taking into account technological developments and impacts on start-ups and SMEs.
  10. It proposes that Member States use existing dedicated channels for communication with SMEs and start-ups, such as ENISA's Computer Security Incident Response Teams and National Data Protection Agencies.
  11. It proposes that the Commission regularly assess certification and compliance costs for SMEs and start-ups, and that penalties take into account their size, interests and economic viability.
  12. It proposes that research and development activity on AI systems not leading to placing on the market or putting into service is not affected by the Regulation, and that freedom of science is respected.

Who is affected

  • Start-ups and SMEs: priority access to regulatory sandboxes, reduced conformity assessment fees, and flexibility in technical documentation.
  • Providers of general purpose AI systems: must work with users to aid compliance and register the system in the EU database.
  • Users of high-risk AI systems: must keep logs for at least six months and assign human oversight to competent persons.
  • Providers of high-risk AI systems: must keep logs for at least six months and provide information to national competent authorities.
  • Member States: should develop initiatives for small-scale providers and use existing communication channels for SMEs and start-ups.

Figures and deadlines

  • Logs must be kept for at least six months unless otherwise provided in Union or national law.
  • Administrative fines of up to 20 000 000 EUR or up to 4 % of total worldwide annual turnover, whichever is higher.
  • SME providers or start-ups supplying incomplete information get 1 month to supply complete information before fines.