On the 14th of June 2023, the European Parliament approved the proposal for a Regulation on Artificial Intelligence – “Artificial Intelligence Act” (hereinafter also only the “AI Act” or the “Regulation”), first proposed on 21.04.2021 by the European Commission .
The Regulation in question aims to ensure that the use and development of AI takes place in conditions of safety, transparency, traceability, and where the privacy and non-discrimination of users is respected.
The Range of Application of the AI Act concerns: Providers who place AI systems on the market or put them into service in the European Union, whether or not they are established in the Union or in a third country; Users of AI-systems located in the European Union; providers and users of AI systems located in a third country, in cases where the output produced by the system is used in the Union.
The new harmonized rules on Artificial Intelligence follow an approach based on risk (so-called “risk-based approach”), wherein more or less stringent obligations for suppliers and system operators are established according to the level of risk that intelligent software and applications are found to pose towards the fundamental rights of users.
The proposed Regulation also classifies as “unacceptable” that AI can be used for certain purposes considered contrary to EU principles and values, such as “social scoring”, i.e. those mechanisms for classifying people based on behavior, socio-economic status or personal characteristics.
The new EU rules provide for an assessment of AI systems based on the different levels of risk they pose. In particular: AI systems classified as unacceptable risk will be banned, as they are systems deemed to be a threat for people; high-risk systems, i.e. those that negatively affect safety or fundamental rights, will be assessed before being put on the market and also throughout their lifecycle and will have to comply with a series of rules – including creating and maintaining an active risk management system and documenting in detail the development of the AI system and its functioning – as well as obligations of transparency towards users; while limited-risk AI systems will still have to comply with the minimum transparency requirements towards users.
Despite the final text still requiring negotiation between the EU Institutions, the approval of the proposed Regulation by the European Parliament, undoubtedly represents an important step forward towards the European regulation of a fundamentally important issue, namely the use of Artificial Intelligence and Generative Artificial Intelligence Systems.