Map of the legislative future of artificial intelligence in Europe (2024-2027) – Health and Medicine

2024-03-13 00:24:14

Following the approval of the European Union AI law, the next steps to be taken in the legislative framework are detailed.

In a context where AI regulation becomes increasingly imperative, the European Union has taken a significant step with the approval of the European Union AI Law (AI Regulation; AI Act, in English). This regulatory framework, planned for gradual implementation from 2024 to 2027, stands as a bulwark in the regulation of AI, establishing a precedent not only in Europe but globally. The next steps in the European legislative framework, as well as their gradual implementation, are detailed below.

Approval of the final text (February-April 2024): during this period, the EU co-legislators will ratify the final text of the AI ​​Law. This phase is crucial, since it consolidates the result of extensive negotiations and adjustments, reflecting the consensus reached between the different Member States and the European institutions. The approval of the final text lays the legal foundations for the future implementation and application of the AI ​​Law. It is a transition period where affected entities must begin to prepare for the upcoming regulatory changes.

Publication in the Official Journal and entry into force (May-June 2024): the AI ​​Law will be published in the Official Journal of the EU, a formal act that marks the beginning of its legal application. It will come into force 20 days after its publication. This short period between publication and entry into force is a legislative standard that allows an orderly transition to the new regulations.

Application of rules on prohibited practices (early 2025): A grace period of six months from the entry into force of the rules on prohibited practices is established. This time frame is essential for entities involved in the development and implementation of AI systems to make the necessary adjustments to comply with the new restrictions. The prohibited practices cover critical areas such as subliminal manipulation and misuse of vulnerabilities of specific groups, highlighting the EU’s commitment to protecting individual rights in the context of AI.

Creation of codes of practice and governance (mid-2025): Nine months after entry into force, the AI ​​Office, in collaboration with the Global Partnership on AI (GPAI), is expected to facilitate the creation of codes of practice. These codes will be essential to guide entities in the ethical and responsible implementation of AI systems. Simultaneously, twelve months after entry into force, rules relating to notifying bodies, AI governance, GPAI and sanctions will apply. This phase emphasizes the importance of an effective governance structure and establishes a framework for accountability and compliance.

Obligations for high-risk systems (mid-2026 and 2027):

• Application for Annex III (mid-2026): high-risk systems listed in Annex III will be subject to specific obligations from mid-2026. This phased approach allows for a more manageable implementation and progressive adaptation by entities affected.
• Application for Annex II (mid-2027): Similarly, for systems that fall under Annex II, the corresponding obligations will begin to apply in mid-2027. This final stage of the schedule provides additional time for adaptation to the more complex regulatory demands associated with these systems.

Implementing the AI ​​Law is a challenge for both developers and regulators. Companies will have to adapt their products and strategies to meet the new standards, which may require significant investments. On the other hand, regulators must ensure effective and equitable application of these standards.

It is essential that Member States’ concerns, such as those expressed by France, are addressed, balancing the promotion of innovation with effective regulation. J.S.LL (GISP-UPC)