The European Union has taken a significant step forward in its regulation of artificial intelligence (AI), as all 27 member states voted to approve the final text of the EU Artificial Intelligence Act. EU Internal Market Commissioner Thierry Breton described the Artificial Intelligence Bill as historic and a world first. The legislation adopts a risk-based approach to regulate the application of artificial intelligence, covering areas such as the government's use of AI in biometric surveillance, regulations for AI systems like ChatGPT, and transparency rules before market access.
The political agreement on the Artificial Intelligence Act was reached in December 2023, and member states have now endorsed the final text. The bill addresses concerns raised by experts about the use of deepfakes, which are fabricated videos created by AI algorithms, posing challenges to truth and fiction in public discourse, especially on social media. The legislation places a particular emphasis on high-risk cases, such as AI used in the classification of job applicants or candidates for educational programs.
France, which had initially opposed the Artificial Intelligence Bill, withdrew its opposition on Friday. Germany also backed the bill after reaching a compromise, and the legislation will now proceed to a vote by a key committee of EU lawmakers on February 13. Subsequently, it will be voted on by the European Parliament in March or April. The implementation of the Artificial Intelligence Act is anticipated in 2026, with specific regulations taking effect earlier. The European Commission is concurrently working to establish an Office for Artificial Intelligence to oversee compliance with high-impact underlying AI models considered to present systemic risks. Additionally, measures have been unveiled to support local AI developers, including the enhancement of the EU supercomputer network for training generative AI models.
