EU delays high-risk AI rules by 16 months in regulatory overhaul
EU governments and European Parliament lawmakers agreed Thursday to delay enforcement of strict rules on high-risk AI systems, marking a significant retreat from the bloc's original timeline for its landmark artificial intelligence regulations.
The tentative agreement, reached after nine hours of negotiations, pushes the implementation deadline for rules governing AI systems used in biometrics, critical infrastructure, and law enforcement to December 2, 2027-16 months later than the previous August 2 deadline.
The changes reflect pressure from businesses that complained about overlapping regulations and administrative costs. Cyprus's deputy minister for European affairs, Marilena Raouna, said the agreement "significantly supports our companies by reducing recurring administrative costs."
What's changing
EU regulators excluded machinery from the AI Act, citing existing sectoral rules that already apply to that sector. This concession directly addresses industry complaints about competing requirements.
The agreement does maintain restrictions on certain AI practices. A ban on creating unauthorized sexually explicit images takes effect December 2, responding to content generated by Elon Musk's xAI chatbot Grok. Mandatory watermarking of AI-generated output will also begin that date.
Dutch lawmaker Kim van Sparrentak said the ban would protect women and girls from "nudifier apps" and deepfakes. "By the end of this year everyone, but especially women and girls will be safe from horrific nudifier apps being widely available on the EU market," she said.
The bigger picture
The EU AI Act entered force in August 2024 and was designed as the world's first comprehensive AI regulation. Even with these changes, officials maintain it remains the strictest AI framework globally.
The agreement requires formal endorsement from EU governments and the European Parliament in coming months. The European Commission initiated the revision as part of a broader effort to simplify digital regulations that businesses said hampered competition with U.S. and Asian companies.
For government professionals, understanding these regulatory shifts is essential. Learn more about AI for Government and how Generative AI and LLM technologies are being regulated across different jurisdictions.
Your membership also unlocks:
