EU Council Agrees on Position to Streamline AI Act Rules
The European Union member states have reached a consensus on a negotiating position aimed at simplifying the implementation of the bloc’s landmark Artificial Intelligence Act. This agreement is part of a broader effort to reduce regulatory burdens and enhance competitiveness across the EU.
Overview of the Agreement
This pivotal agreement, forged by the Council of the European Union, is included in the EU’s “Omnibus VII” legislative package. The goal is to streamline digital legislation and make compliance with AI rules more practical for businesses while still maintaining essential safeguards.
Officials have indicated that this proposal is designed to provide greater legal certainty and ensure more consistent enforcement of AI rules across member states. As remarked by Marilena Raouna, Deputy Minister for European Affairs of Cyprus, these changes will foster innovation while enabling companies to adapt more effectively to the regulatory environment.
Key Changes in the Revised Plan
One significant alteration pertains to the timeline for applying certain rules to high-risk AI systems. The Council proposes the following:
- Requirements for standalone high-risk systems will take effect on Dec. 2, 2027.
- Rules for high-risk systems embedded in products will commence on Aug. 2, 2028.
This delay aims to ensure that technical standards and compliance tools are available before enforcement begins, thereby aiding businesses in their transition.
Strengthening Protections Against Harmful AI Uses
The Council has also introduced new provisions aimed at enhancing protections against harmful uses of AI technology. Key measures include:
- An explicit ban on AI systems that generate non-consensual sexual or intimate imagery and child sexual abuse material.
Furthermore, amendments have been made to restore requirements for AI providers to register systems in the EU database for high-risk applications, even if they believe their systems qualify for exemptions. The proposal continues to uphold strict conditions regarding the processing of sensitive personal data to detect and correct bias in AI systems.
Other Regulatory Adjustments
Further changes include:
- The establishment of national AI regulatory sandboxes has been delayed until December 2027.
- Clarification of the supervisory role of the EU’s AI Office while maintaining certain responsibilities for national authorities in areas such as law enforcement and financial regulation.
Next Steps
Negotiations between the Council and the European Parliament are expected to commence shortly, as stakeholders seek to finalize the framework for the EU’s AI regulations.
Assistance Available
If there are any questions or concerns regarding global guidelines, regulations, or laws, individuals are encouraged to reach out to BABL AI. Their audit experts can provide valuable insights and ensure compliance with the new regulations.
