EU Commission proposes legal framework for AI

May 5, 2021 Benn-Ibler Rechtsanwälte

The European Commission introduced a proposal for a regulation on laying down harmonized rules on artificial intelligence (Artifical Intelligence Act). This proposal represents the European Commission’s first ever legal framework for AI.

Below, the core elements of the proposal are listed:

• The proposal follows a risk-based approach. The proposal differentiates between systems of acceptable, high and low risk. Consequently, this implies a list of prohibited AI systems which are considered unacceptable, as they contravene certain EU values.
• Special regulations are intended for AI systems that pose a high risk to the health, safety or fundamental rights of persons. These regulations are to be based on the intended use of the AI system in accordance with existing legislation on product safety.
• Transparency requirements are foreseen for systems that interact with humans (e.g. chatbots) or that are used for emotion recognition or biometric categorisation, or to generate or manipulate content (e.g. ‘deep fake’).
• The European Artificial Intelligence Board, composed of representatives from the Member States and the Commission, is to facilitate a harmonized implementation of the regulation
• National authorities (e.g. national supervisory authorities) are to supervise the application and implementation of the regulation.
• Providers of AI systems are obligated to report and investigate AI-related incidents and malfunctioning.
• The proposal provides for three levels of administrative fines: Up to 10, 20 or 30 million EUR or, if the offender is a company, up to 2%, 4% or 6% of its total worldwide yearly annual turnover for the preceding financial year, whichever is higher.

The proposal will be reviewed and evaluated by the European Council and the European Parliament.

European Commission (21.04.2021), COM (2021) 206 final, 2021/0106 (COD).