The Council gives the green light to the digital omnibus package on AI: simplification of the Artificial Intelligence Regulation

On June 29, 2026, the Council gave its final approval to the regulation amending the AI Regulation (Regulation (EU) 2024/1689) to simplify and streamline its application, as part of the simplification package 'Omnibus VII (digital)'. This concludes the legislative procedure: the Parliament had already approved the text on June 16, 2026, and the Council's approval has been the final step for its formal adoption. It will enter into force three days after its official publication, which is expected in the coming days.

Key Changes

• Postponement of key obligations related to high-risk AI systems, while keeping the architecture and risk-based approach of the AI Regulation intact.
• Prohibition of 'nudifier' applications: AI systems generating child sexual abuse material (CSAM) or creating images, videos, or audio depicting the intimate parts of an identifiable person or sexually explicit activities without their consent are prohibited unless they incorporate adequate technical safeguards.
• Reduction of overlaps in industrial AI: products already regulated by specific sector legislation (medical devices, toys, lifts, machinery, vessels) are exempt from double regulation; Regulation (EU) 2023/1230 on machinery is exempt from the direct application of the AI Regulation, and the Commission will incorporate health and safety requirements for high-risk AI systems through delegated acts.
• A more precise definition of 'safety component': products with AI functions that merely assist the user or optimise performance will not automatically be classified as high-risk if their failure does not pose risks to health or safety.
• Processing of personal data strictly necessary to detect and correct biases, with safeguards, both in high-risk systems and in low-risk systems.
• Support for growing businesses: exemptions for SMEs extend to small mid-cap enterprises (SMEs).
• Restoration of the simplified registration process for systems that the provider considers exempt from high-risk classification.

Deadlines

• 2 December 2026: new deadline for labelling and transparency obligations regarding AI-generated content for systems already on the market before 2 August 2026, which have a four-month transition period to comply.
• 2 December 2026: deadline for companies to align their systems with the prohibition of 'nudifying' applications.
• 2 August 2027: new deadline for national authorities to establish regulatory sandboxes.
• 2 December 2027: entry into force of obligations for high-risk autonomous systems (including biometrics, critical infrastructure, education, employment, essential services, justice, law enforcement, and border management).
• 2 August 2028: application to AI systems integrated as safety components in products subject to specific safety legislation and market oversight in the sector.

Other Issues

• Deepfakes and synthetic content: AI-generated content must bear a label in a machine-readable format to improve transparency and enable detection and traceability, with the new application date set for 2 December 2026.
• AI Office: its exclusive competence to oversee general-purpose AI model-based systems is reinforced and clarified when the model and system come from the same provider, or from providers within the same company, as well as systems that constitute or are integrated into large platforms or search engines (VLOP/VLOSE).
• Exceptions in favour of national authorities: these retain jurisdiction in areas such as law enforcement, border management, financial institutions, and administration of justice, as well as systems linked to products in Annex I.

Key Recommendations

• Review the internal compliance timeline: the postponement of high-risk norms provides additional leeway, but it is advisable to take advantage of the greater certainty offered by fixed deadlines to plan for compliance.
• Check exposure to the prohibition of 'nudifiers' and labelling obligations, whose deadlines are already set for 2 December 2026.

Informative note from the TMT department of ECIJA Madrid.