EU reaches tentative deal on simpler AI rules
Key Points:
- EU member states and European Parliament negotiators have reached a provisional deal to delay implementation of key rules under the EU's Artificial Intelligence Act, pushing back obligations for high-risk AI systems to December 2, 2027, from the original August 2, 2024 deadline.
- The agreement excludes machinery from the AI Act, allowing such systems to comply only with existing sectoral safety regulations, a move seen by critics as yielding to business pressures.
- A mandatory watermarking requirement for AI-generated content will take effect from December 2, 2024, while a ban on AI applications that create unauthorized sexually explicit deepfake content, including child sexual abuse material, will also be enforced by the same date.
- The amendments aim to reduce administrative burdens and legal uncertainties for companies, supporting smoother implementation and strengthening the EU's digital sovereignty and competitiveness, according to Cyprus’ Deputy Minister for European Affairs.
- These changes are part of the European Commission’s broader effort to simplify digital regulations, with the provisional agreement pending formal approval by the European Parliament and EU governments.