The Council and Parliament have reached a groundbreaking agreement on the world’s first rules for Artificial Intelligence (AI) after three days of intense negotiations. The provisional agreement, known as the Artificial Intelligence Act, aims to establish harmonized rules for AI to ensure safety, respect fundamental rights, and uphold EU values in AI systems placed on the European market and used within the EU. This landmark legislation is designed not only to provide a regulatory framework for AI but also to promote investment and innovation in AI across Europe.
Carme Artigas, the Spanish Secretary of State for Digitalisation and Artificial Intelligence, describes the agreement as a historic achievement and a significant step toward addressing global challenges in the rapidly evolving technological landscape. The delicate balance achieved in the negotiations emphasizes boosting innovation and AI adoption across Europe while respecting the fundamental rights of citizens.
The AI Act is a flagship legislative initiative with the potential to foster the development and adoption of safe and trustworthy AI throughout the EU’s single market. It adopts a ‘risk-based’ approach, wherein the severity of the risk determines the stringency of the rules. As the first legislative proposal of its kind globally, the AI Act aims to set a global standard for AI regulation, similar to the impact of the General Data Protection Regulation (GDPR) on data protection.
The key elements of the provisional agreement include:
- Rules on high-impact general-purpose AI models and high-risk AI systems.
- Revised governance system with EU-level enforcement powers.
- Extension of prohibitions with safeguards for remote biometric identification.
- Enhanced protection of rights through mandatory fundamental rights impact assessments for deployers of high-risk AI systems.
The agreement clarifies definitions and scope by aligning with the OECD’s approach, ensuring clear criteria for distinguishing AI from simpler software systems. It also specifies that the regulation does not apply to areas outside EU law, military or defense purposes, research and innovation, or non-professional use of AI.
The classification of AI systems into high-risk and prohibited practices is a crucial aspect of the agreement. It introduces a horizontal layer of protection, ensuring that AI systems posing limited risks are subject to lighter transparency obligations, while high-risk systems face more stringent requirements to access the EU market.
Law enforcement exceptions are considered, balancing the need for AI use in vital law enforcement work with safeguards against potential misuse. Provisions for real-time remote biometric identification in public spaces are strictly defined, limiting use to specific law enforcement purposes with additional safeguards.
The agreement introduces specific rules for general-purpose AI systems and foundation models, addressing situations where AI systems are used for various purposes. A new governance architecture is established, including an AI Office, a scientific panel, an AI Board, and an advisory forum for stakeholders.
Penalties for AI Act violations are set as a percentage of the offending company’s global annual turnover, with proportionate caps for SMEs and start-ups. The agreement emphasizes transparency, fundamental rights impact assessments, and increased transparency regarding the use of high-risk AI systems.
To support innovation, the agreement modifies provisions for AI regulatory sandboxes, allowing testing in real-world conditions. Specific measures are included to reduce administrative burdens for smaller companies.
The AI Act is expected to come into force two years after its entry into force, with ongoing technical work to finalize details before formal adoption by the co-legislators. The agreement builds on the Commission’s proposal presented in April 2021, aligning with the EU’s policy to foster safe and lawful AI development and uptake across the single market. The proposal follows a risk-based approach, providing a horizontal legal framework for AI to ensure legal certainty and promote investment and innovation in
