In April 2021, the Commission proposed a regulation to lay down harmonised rules on artificial intelligence (AI Act). Since then, the Council and the Parliament have worked on their positions to start the negotiations. In December 2022, the Council adopted its common position (‘general approach’). On the other hand, the Parliament has yet to agree on its text, and seven Committees are working on this new regulation. Five of them (ITRE ENVI, CULT, TRAN, and JURI Committee) have already published their opinions, and the Joint Committee Responsible (IMCO and LIBE) is still working on the report.
SCOPE
The rules will apply to all providers placing on the market or putting into service AI systems in the Union, irrespective of whether those providers are established within the Union or in a third country. The proposal significantly clarifies that it applies to providers and users of AI systems that are located in a third country, where the output produced by the system is used in the Union. Even though the text is under discussion, none of the Parliament's Committees seems wanting to change this.
DEFINITION and CLASSIFICATIONS
The Act's key is to define i) what AI systems are and to ii) what products the regulation would apply. Currently, the discussions in the Parliament led to the alignment with the OECD definition. Among other definitions and amendments, another crucial agreement to find is the one related to the classification of high-risk AI systems. They would have a particular set of rules and subcategories rules, such as the new EU database for stand-alone high-risk AI systems. Furthermore, it is explained below how these definitions would be relevant to other EU regulations.
SPECIAL MEASURES and IMPLEMENTATION
Currently, there is a general agreement in the Parliament, and also in the Council, related to the need for special measures to support innovation and small-scale providers and users.
In addition, the proposal established the European Artificial Intelligence Board to assist in the in implementing the whole Act.
AI LIABILITY DIRECTIVE
In September 2022, the Commission proposed the AI Liability Directive. The directive would only apply to non-contractual, fault-based civil law claims for damages. The proposal addresses two major topics:
Next Steps
In the case of the AI Act, the Parliament’s final position is yet to come, but discussions are in progress. Once MEPs agree on that, the Council, the Commission and the Parliament will negotiate in trilogues the final text.
The AI Liability Directive is at an early stage. There isn’t any document from the Parliament yet, and the discussions in the Council are just starting.
Comentarios