- within Insurance and Tax topic(s)
For further information on any of the issues discussed in this publication please contact the related contact(s) on this page.
Background
The EU AI Act, the first-ever legal framework on AI, entered into force in August 2024 with the objective of setting down harmonised rules on artificial intelligence to address associated risks and to foster trustworthy AI in the EU.
While its implementation was staggered with some rules already applicable, the majority of the EU AI Act’s provisions were due to apply from 2 August 2026.
European Commission Digital Package on Simplification
As part of its broader simplification agenda to boost competitiveness in the EU, the European Commission published a Digital Omnibus Package in November 2025. Amongst its proposals were a series of targeted amendments to the EU AI Act aiming to simplify its provisions to make its implementation for EU businesses easier while maintaining its intended level of protection.
What has been agreed by the EU rule-makers?
On 7 May 2026, following negotiations with the European Parliament and the Council of the EU, the European Commission announced that political agreement had been reached between the institutions on targeted changes to the EU AI Act.
While the full text of the proposed amendments has not yet been published, the press releases from the three institutions indicate that these changes include:
-
delaying the application date for complying with the rules governing high-risk AI systems classified under Annex III of the AI Act by 16 months until 2 December 2027. This includes AI systems used in employment and HR amongst others. This is intended to allow adequate time for technical standards and other support tools such as guidelines for implementation to be published by the European Commission before the rules start to apply;
-
the extension of certain exemptions previously available only to SMEs now being made available to “small mid-cap” companies;
-
strengthening the enforcement powers of the European Commission’s AI Office to support supervision of certain AI systems, including those built on general-purpose models and those embedded in very large online platforms and very large search engines;
-
delaying the deadline for complying with transparency obligations for artificially generated content known as watermarking obligations which allow for the detection and tracing of AI-generated content from 2 August 2026 to 2 December 2026. Importantly, all other transparency obligations imposed under the EU AI Act will retain the existing application date of 2 August 2026 meaning that providers and, where relevant, deployers of AI systems should continue to prioritise compliance with these obligations ahead of the upcoming deadline; and
-
delaying the obligation for Member States to establish at least one national AI regulatory sandbox until 2 August 2027.
What next?
The political agreement reached by the co-legislators must now be formally adopted by the European Parliament and the Council of the EU. The co-legislators expect to do so before 2 August 2026, the current application date for rules on high-risk AI systems.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
[View Source]
