In a significant move, European Union member states have given their nod to the final text of the EU’s AI Act, marking a crucial step forward in the regulatory landscape for artificial intelligence (AI). The approval, granted on February 2, was secured after France and Germany withdrew their objections to the AI Act.
Thierry Breton, the Commissioner for Internal Market of the EU, confirmed the endorsement of the political agreement reached in December 2023 by all 27 member states. In a post on social media platform X, Breton emphasized the historic nature of the AI Act, hailing it as a global first.
The AI Act is a risk-based strategy designed to regulate AI applications comprehensively. The agreement encompasses the governmental use of AI in biometric surveillance, regulations for AI systems like ChatGPT, and transparency rules that must be adhered to before entering the market.
Following the political agreement in December, efforts were initiated to translate the agreed-upon positions into a final compromise text for approval by lawmakers. This process concluded with the “coreper” vote on February 2, where the permanent representatives of all member states cast their votes.
Experts have expressed concerns about deepfakes—realistic yet fabricated videos created by AI algorithms—appearing on social media platforms and blurring the line between truth and fiction in public discourse.
Margrethe Vestager, the Executive Vice President of the European Commission for A Europe Fit for the Digital Age, highlighted the significance of Friday’s agreement as a crucial step toward implementing the AI Act. She emphasized the core principle that the riskier the AI, the greater the liabilities for developers, particularly in high-risk cases such as sorting job applicants or gaining admission to educational programs.
The agreement on Friday coincided with France withdrawing its objection to the AI Act, and on January 30, Germany also endorsed the act after reaching a compromise, as confirmed by the Federal Minister for Digital Affairs and Transport, Volker Wissing.
The AI Act is now set to progress towards legislation, with a vote by a crucial EU lawmaker committee scheduled for February 13, followed by a European Parliament vote in March or April. The implementation is expected in 2026, with specific provisions taking effect earlier.
To ensure compliance with a group of high-impact foundational models considered to have systemic risk, the European Commission is establishing an AI Office. Additionally, measures have been unveiled to support local AI developers, including the upgrading of the EU’s supercomputer network for generative AI model training.
