EU Countries Approve Final Text of Landmark Artificial Intelligence Act, Set to Enforce by 2026
Summary:
The European Union's member countries have approved the final text of the EU's Artificial Intelligence Act, representing a significant forward step in AI regulation, including models like Microsoft-supported ChatGPT. The landmark Act covers governmental use of AI in biometric surveillance, setting standards for AI systems and establishing transparency rules. The agreement was reached after the withdrawal of objections by France and Germany, paving the way for a vote by an EU lawmaker committee and the European Parliament in the coming months. The Act is expected to be enforced by 2026, and the Commission is setting up an AI Office to ensure compliance and assist local developers.
The steps toward setting up regulations for artificial intelligence (AI) applications, including Microsoft-supported models like ChatGPT, have seen significant progress in Europe, with European Union (EU) countries supporting the final version of the AI Act. The EU's Industry Commissioner, Thierry Breton, announced that the agreement reached politically in December got the green light from all the 27 member states and declared the AI Act a historical first worldwide. The AI Act, a risk-assessment approach to AI regulation, encompasses governmental use of AI in biometric surveillance, establishing norms for AI systems such as ChatGPT, and setting transparency standards prior to their market introduction.
After reaching a political agreement in December, the process was initiated to translate the agreed upon points into a compromise text finalised by lawmakers. The Coreper I Ambassadors stamped their approval on the drafted rules on February 2. Experts have raised concerns about deepfakes - fabricated yet convincingly real videos generated by AI algorithms that have access to extensive online footage - as they blur the line between real and false in public conversation.
EU's digital chief Margrethe Vestager highlighted that the agreement was a notable stride toward the AI Act, basing it on the principle that the risks associated with AI directly influence responsibilities for developers. For instance, if AI is used for job applicant screening or college admission, the AI Act addresses high-risk instances specifically. France, the last EU state to oppose the AI Act, retracted its resistance on February 2, making the agreement feasible. On January 30, Germany, following initial resistance, also supported the AI Act after the Federal Minister for Digital Affairs and Transport, Volker Wissing, announced that a compromise was reached.
The progress of the AI Act to legislation is set to proceed with an EU lawmaker committee voting on February 13, followed by a vote in the European Parliament tentatively scheduled for March or April. The enforcement of the Act is anticipated by 2026, with specific rules coming into effect sooner. Measures are being taken by the Commission to instate an AI Office in charge of ensuring adherence to high-impact foundational models that could pose systemic risks. Additionally, regulations have been laid out to help local AI developers including an upgrade for the EU's supercomputer network aiding the training phase of generative AI models.
Published At
2/3/2024 10:59:35 AM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.