EU AI Act Amendments Approved by European Parliament

What the AI Omnibus Changes Mean for Businesses Preparing for EU AI Act Compliance

On 19 November 2025 the European Commission published proposals for a Digital Omnibus on AI (“AI Omnibus”) to address specific implementation challenges and lighten regulatory burdens that were identified following the entry into force of the EU AI Act . However, due to deadlines under the EU AI Act that were due to take effect in August 2026, the AI Omnibus had to be fast-tracked.

On 7 May 2026, the European Parliament and the European Council reached a political agreement on the AI Omnibus. However, this still needed to be formally approved by both before it becomes law.

On 16 June 2026, the European Parliament endorsed the amendments to the EU AI Act which are part of the AI Omnibus.

The AI Omnibus does not change the overall structure of the EU AI Act or its risk-based approach. Instead, it will introduce a number of amendments such as postponing certain parts of the EU AI Act in order to ensure that the necessary standards and support measures were in place. The changes included:

• the obligations on high-risk AI systems will now apply from 2 December 2027 for stand-alone high risk AI systems and 2 August 2028 for high-risk AI systems embedded into products;
• the obligations to mark AI-generated content will be delayed until 2 December 2026 for AI systems placed on the market before 2 August 2026;
• a prohibition on AI systems that generate non-consensual intimate or sexual content as well as child sexual abuse material will take effect from 2 December 2026. Providers will not be allowed to place these systems on the EU market, unless they include adequate technical safeguards to prevent the creation of such material. This prohibition will also apply to deployers using AI systems to create such content;
• a simplified and proportionate registration obligation for non-high-risk AI systems;
• a strict necessity standard for the processing of special categories of personal data for the purpose of ensuring bias detection;
• removing AI-enabled machinery products from the direct applicability of the EU AI Act’s high-risk regime and clarifying that they need only comply with sectoral approach in the Machinery Regulation. The European Commission will however be required to adopt delegated acts amending Annex III of the Machinery Regulation in order to include health and safety requirements relating to AI systems that are classified as high-risk and that are either safety components in machinery or machinery themselves.

Now the AI Omnibus will pass to the European Council who is expected to formally endorse the agreed text. Once this happens, the text will undergo a legal/linguistic revision before being formal adopted by the co-legislators.

While these final steps may appear procedural, businesses should not treat this as a distant development. Instead, this is a critical window to assess exposure, strengthen governance frameworks and ensure compliance readiness. Organisations that act early will be best placed not only to mitigate regulatory risk but also to leverage AI responsibly as a strategic advantage in an increasingly regulated landscape.

For further information and support, please contact Maureen Daly at mdaly@reddycharlton.ie or your usual contact in Reddy Charlton LLP