France and Germany withdrew their objections to the AI Act concurrently with the adoption on February 2. Member states have voted to accept the final language of the EU’s AI Act, advancing the legislative framework for artificial intelligence (AI) within the EU. Thierry Breton, EU Commissioner for Internal Markets, verified that all 27 member states “endorse the political agreement reached in December” of 2023. He said that the AI Act is a first for the globe and historical in a post on the social networking site X. A risk-based approach to regulating AI uses is the AI Act.
The agreement addresses the use of AI by the government for biometric monitoring, the regulation of AI platforms such as ChatGPT, and the need for transparency prior to market entry.
The December political agreement marked the beginning of the process of turning agreed-upon positions into a final compromise language for legislative approval. This process culminated on February 2 with the vote of the permanent representatives of all member states, known as the “coreper” vote.
Experts have expressed serious concerns about deepfakes, which are artificial intelligence (AI) algorithms trained on internet footage to make realistic-looking but fake movies that emerge on social media and blur the distinction between fact and fiction in public debate.
Margrethe Vestager, Executive Vice President of the European Commission for A Europe Fit for the Digital Age, stated that the agreement reached on Friday represents a major advancement towards the AI Act. She declared:
“Founded on the straightforward tenet that developers have larger liabilities the riskier the AI. For instance, if employed to rank candidates for employment or admission to a school. Because of this, the #AI Act concentrates on high-risk scenarios.
On Friday, the accord was reached when France decided not to oppose to the AI Act. On January 30, Germany also supported the measure following Volker Wissing’s announcement that a compromise had been struck as Federal Minister for Digital Affairs and Transport.
A significant EU legislative committee is scheduled to vote on the AI Act on February 13; the European Parliament will then vote on the bill in March or April, bringing it closer to becoming law. It is anticipated that it will be implemented in 2026, with some aspects going into force sooner. The European Commission is moving to create an AI Office in order to oversee adherence to a set of influential foundational models that are thought to pose a systemic risk. It also revealed plans to help regional AI developers, such modernising the EU supercomputer network to train generative AI models.