The European Artificial Intelligence Regulation (so called « Al Act ») was voted on March 13th, 2024 by the European Parliament in plenary session. The aim of the EU legislator is to ensure Al in Europe is safe, respects fundamental rights and democracy, while businesses can thrive and expand.

Read below our overview of the practical impacts for companies creating or using Al systems, the first in our series dedicated to the Al Act.

SCOPE AND TIMELINE

• The Al Act will apply to both public and private actors located within and outside of the EU as long as the Al system is placed on the Union market or its use affects people located in the EU;
• Al Act obligations will not apply to:
  • research, development and prototyping activities preceding the release on the market;
  • Al systems that are exclusively for military, defense or national security purposes, regardless of the type of entity carrying out those activities;
  • providers of free and open-source models, except for providers of general purpose Al models with systemic risks.
• Once the final version of the Al Act is released, supposedly by the end of April 2024, and published in the Official Journal of the EU; companies will have to comply within a period of 6 to 36 months, depending on the requirements applicable to them

ENFORCEMENT AND SANCTIONS

• Member States must designate national Al authorities to supervise the application
  and implementation of the Regulation. These authorities will be responsible for
  representing the country in the European Artificial Intelligence Board whose role
  is to facilitate the application of the Regulation in Europe;
• Member States may impose penalties of up to €35 million or 7% of the total worldwide annual turnover of the preceding financial year.

OVERVIEW OF THE AI ACT RISK-BASED APPROACH

WHERE TO START ?