Timeline

The European Parliament passes the AI Act

Adopted 523 votes to 46 after three years of negotiation; the Council's formal sign-off and publication in the Official Journal followed months later.

  • Government & policy
  • Major

The European Parliament voted 523 to 46, with 49 abstentions, to adopt the Artificial Intelligence Act, ending roughly three years of negotiation since the European Commission’s original 2021 proposal. The vote made law a risk-tiered framework that had been the most closely watched attempt anywhere to regulate AI comprehensively rather than sector by sector.

The Act bans a specific list of uses outright: biometric categorisation based on sensitive characteristics such as race or sexual orientation, untargeted scraping of facial images to build recognition databases, emotion-recognition in workplaces and schools, social scoring, and AI systems designed to manipulate behaviour or exploit vulnerabilities. Systems classed as “high-risk” — those used in critical infrastructure, employment, education, essential services, law enforcement or the administration of justice — must meet obligations on risk assessment, data quality, logging, transparency and human oversight before they can be placed on the EU market. General-purpose AI models face a separate layer of requirements: technical documentation, compliance with EU copyright law, and published summaries of training data, with additional obligations — model evaluation, systemic-risk assessment, incident reporting — for the most capable systems, defined by a training-compute threshold.

Parliamentary adoption was not final enactment. The text still required the Council of the EU’s formal sign-off, publication in the Official Journal, and a staged timeline before individual provisions took effect — the banned-practices rules first, general-purpose-model obligations later, and most high-risk-system obligations only after a longer transition. The Council gave that formal sign-off in May 2024, and the Act was published and entered into force that summer.

Because most large AI developers sell into the EU market, the Act’s extraterritorial reach made it a de facto floor for global compliance practice in ways aimed purely at national markets rarely achieve, and it became the reference point against which US and UK “lighter-touch” approaches were repeatedly measured over the following two years.

Referenced by