The ambassadors of the 27 countries of the European Union have unanimously approved the world’s first comprehensive rulebook for Artificial Intelligence (AI). This approval comes after a political agreement was reached in December on the main points of the AI Act, a bill to regulate AI based on its potential harm. The agreement faced opposition, particularly from France, Germany, and Italy, who wanted a lighter regulatory regime for powerful AI models. However, a compromise was reached with a tiered approach, imposing transparency rules for all models and additional obligations for models posing systemic risks. The implementation of the AI Act will now be closely watched by member states, who still have room to influence its execution through secondary legislation.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| EU approves comprehensive rulebook for Artificial Intelligence | Regulation of AI models | AI models subject to transparency rules and additional obligations for high-risk models | Concerns about the potential harm caused by AI models |
| Debate over regulation of powerful AI models | Disagreement over regulatory regime for powerful AI models | Compromise reached with horizontal transparency rules for all models and additional obligations for high-risk models | Balancing the regulation of powerful AI models while supporting European start-ups |
| Negotiating agreement on law enforcement chapter of AI Act | Disagreement over law enforcement exceptions | Final agreement reached after intense negotiations | Balancing law enforcement requirements with the regulation of AI |
| Resistance from France to the provisional agreement | France opposes AI rulebook | France eventually agrees to support the text with conditions | Concerns about the impact on AI development and trade secrets |
| EU countries seek clarifications and raise concerns about AI Act | Countries seek clarification on critical terms and express concerns about data protection and consumer law | Implementation of the AI law to address concerns and ensure international alignment | Ensuring the balance between transparency and protection, and addressing privacy and consumer protection concerns |
| Establishment of the European Artificial Intelligence Office | Creation of AI Office | AI Office responsible for overseeing AI models | Need for a centralized authority to manage AI regulation |
| Adoption and implementation timeline of the AI Act | AI Act adoption and implementation process | Full adoption and enforcement of the AI Act, with gradual application of different rules | Incremental implementation to allow for preparation and adjustment |
| Concerns about the independence of Hungary’s digital authority | Concerns about lack of independence in enforcing Digital Services Act | MEPs express concerns about the authority’s independence | Ensuring effective enforcement of digital regulations |