• The European Union’s AI Act, aimed at regulating artificial intelligence applications, passes a crucial vote by Member State representatives.
• The regulation outlines prohibited uses of AI, governance rules for high-risk applications, and transparency requirements for AI chatbots.
• All 27 EU Member States unanimously back the final text, overcoming opposition primarily led by France.
• The European Parliament will now vote on the compromise text, with adoption expected in the coming months.
• The Act will come into force gradually, with banned uses of AI taking effect after six months and rules on foundational models after a year.

After months of negotiations and political wrangling, the European Union’s AI Act has overcome its last significant obstacle. Member State representatives voted to confirm the final text of the draft law, marking a critical step towards adoption.

Following a political agreement reached in December, EU co-legislators engaged in exhaustive negotiations to finalize the text. Today’s Coreper vote, affirming the draft rules, signals a crucial milestone in the regulatory journey of AI in the EU.

The AI Act delineates prohibited uses of AI, emphasizing risk mitigation in applications that could harm health, safety, fundamental rights, and democracy. It also imposes transparency requirements on AI chatbots. However, ‘low-risk’ AI applications fall outside the law’s scope.

Despite initial opposition, all 27 ambassadors of EU Member States unanimously endorsed the final text. This resounding support alleviates concerns about the Act’s derailment and ensures its progression towards adoption.

The baton now moves to the European Parliament, where lawmakers will conduct a final vote on the compromise text. Given the widespread backing and minimal dissenting voices, adoption is expected in the coming months.

Once adopted, the Act will come into force gradually. Banned uses of AI will take effect after six months, with rules on foundational models following after a year. The bulk of the regulations will apply two years after publication.

The European Commission is already laying the groundwork for implementation, including the establishment of an AI Office to oversee compliance. Additionally, measures to support homegrown AI development indicate the EU’s commitment to fostering innovation while ensuring responsible AI deployment.