The Council of the European Union and the European Parliament reached a provisional agreement on Friday on the groundbreaking proposal for the Artificial Intelligence Act (AI Act), marking the world’s first comprehensive set of rules governing artificial intelligence.

The primary objective of the AI Act is to regulate AI based on its potential risks to society, employing a risk-based approach where stricter rules apply to higher-risk AI systems. This landmark legislation has the potential to set a global standard for AI regulation, similar to the impact of the EU’s General Data Protection Regulation (GDPR) in the field of data protection.

Key provisions of the provisional agreement include rules on high-impact general-purpose AI models that can pose systemic risks in the future, as well as regulations for high-risk AI systems. The governance framework was revised, granting enforcement powers at the EU level. The list of prohibitions has been extended, while allowing for the use of remote biometric identification by law enforcement authorities in public spaces, subject to safeguards. To protect fundamental rights, deployers of high-risk AI systems will be required to conduct a fundamental rights impact assessment before deploying such systems.

The European Commissioner for Internal Market, Thierry Breton, hailed the agreement as a historic achievement. He emphasized that the EU has become the first continent to establish comprehensive rules for the use of AI. Breton stated, “The AI Act is much more than a rulebook—it’s a launchpad for EU startups and researchers to lead the global AI race.” Roberta Metsola, President of the European Parliament, welcomed the development, describing it as a pivotal moment for digital Europe, saying that it sets global standards.

However, the AI Act has not been without its critics. In an open letter signed by over 150 AI professors and PhDs, scholars called for exemptions for research and open-source projects from burdensome requirements when sharing foundation models. Antti Honkela, a professor of data science at the University of Helsinki, voiced concerns about the potential impact on European AI research. He emphasized the importance of allowing the sharing of models, including powerful foundation models, which he deemed vital for AI research. He urged policymakers to carefully consider the implications of these provisions to avoid stifling innovation in European AI research.