EU AI Digital Omnibus: A Closer Look at the Proposed Amendments to the EU AI Act

SUMMARY

On 19 November 2025, the European Commission published its Digital Omnibus on AI, a legislative proposal to streamline the implementation of the EU AI Act. The proposal addresses delays in harmonized standards and official guidance while adjusting compliance timelines and reducing administrative burdens for businesses.

Key Points

• Deadline Extensions: High-risk AI compliance is now tied to the availability of standards, with "backstop" dates extending preparation by up to 16 months.
• Centralized Oversight: The EU AI Office gains exclusive enforcement powers over general-purpose AI and systems integrated into very large online platforms.
• Small and Medium Enterprises (SME) & Mid-Cap Relief: Regulatory benefits are extended to "small mid-caps," and quality management requirements are simplified for all SMEs.
• Reduced Red Tape: Registration requirements are eliminated for AI systems deemed "not high-risk" under Article 6(3).
• Data Flexibility: A new legal basis allows providers to process sensitive data for bias detection and correction across all AI risk levels.

WHAT DOES THE AI DIGITAL OMNIBUS ENTAIL?

The Digital Omnibus on AI is a legislative proposal that amends specific provisions of the EU Artificial Intelligence Act. The term "omnibus" refers to a legislative tool used to amend multiple provisions simultaneously.

The AI Act, adopted as Regulation (EU) 2024/1689, established a comprehensive regulatory framework for artificial intelligence. Industry feedback revealed implementation challenges: businesses faced uncertainty about risk classifications, and essential support infrastructure, including harmonized standards and official guidance, remained incomplete.

In several member states, competent authorities responsible for enforcement had yet to be designated.

The Digital Omnibus addresses these issues through targeted amendments that adjust timelines, clarify provisions, and reduce administrative burdens.

HOW ARE COMPLIANCE SCHEDULES ALTERED FOR HIGH-RISK AI SYSTEMS?

The proposal links the application of high-risk requirements to the availability of supporting infrastructure. Rules now take effect only after the Commission confirms the completion of relevant standards and guidance, followed by a transition period of six to twelve months.

If standards development faces delays, backstop dates apply:

• 2 December 2027: For standalone high-risk use cases (e.g., biometrics, employment).
• 2 August 2028: For AI embedded in regulated products (e.g., medical devices).

This provides up to 16 months of additional preparation time compared to the original Act.

WHAT ARE THE EFFECTS ON AI SYSTEMS CURRENTLY AVAILABLE?

• Legacy Systems: Providers of high-risk systems lawfully on the market before the compliance date may continue selling identical units without retrofitting. However, systems used by public authorities must reach full compliance by 2 August 2030.
• Transparency: For systems generating synthetic content, marking and watermarking obligations for existing models will now apply from 2 February 2027.

ARE REGISTRATION DUTIES BEING MODIFIED?

Yes. The proposal eliminates the requirement to register AI systems in the EU database if they are deemed "not high-risk" under Article 6(3). While providers are no longer required to register these systems publicly, they must still document their risk assessments and provide them to authorities upon request.

WHAT SUPPORT IS OFFERED TO SMALLER BUSINESSES?

The Digital Omnibus extends SME benefits to small mid-cap companies (SMCs) (entities with fewer than 750 employees or €150M turnover). Additionally, simplified quality management system requirements are expanded to all SMEs, not just microenterprises, allowing for compliance measures proportionate to organization size.

HOW ARE CONFORMITY EVALUATIONS DEFINED MORE CLEARLY?

For AI systems that fall under both "standalone" and "embedded" categories, the proposal clarifies that providers must follow the assessment procedures applicable to AI embedded in regulated products. It also extends Article 60 testing opportunities, allowing providers to conduct controlled trials before full certification.

HOW IS ENFORCEMENT ADJUSTED IN THE PROPOSAL?

The AI Office gains exclusive enforcement powers over:

• General-purpose AI (GPAI) models and systems developed by the same provider.
• AI integrated into Very Large Online Platforms (VLOPs) or search engines under the Digital Services Act.

This shifts oversight from individual Member States to centralized EU oversight for the most significant AI systems, a development closely monitored by AI law firm and lawyers advising multinational technology companies and digital platforms.

WHAT MODIFICATIONS APPLY TO AI EDUCATION AND BIAS CORRECTION?

• AI Literacy: The proposal removes the direct obligation for providers to ensure staff literacy, instead requiring the Commission and Member States to encourage such education.
• Bias Correction: The permission to process "special category data" (e.g., race, health) for bias detection is expanded to all AI systems, not just high-risk ones, facilitating fairer development across the industry.

WHAT ISSUES DOES THE DIGITAL OMNIBUS NOT RESOLVE?

Several critical industry concerns remain:

• Definitional Clarity: Key terms like "provider" remain ambiguous, with legal certainty relegated to non-binding guidelines.
• Redundancy: High-risk AI still requires a "Fundamental Rights Impact Assessment," which often overlaps with existing GDPR data protection assessments.
• Research Exemptions: The exemption for scientific research remains narrow, potentially excluding real-world testing for commercial R&D in fields like pharmaceuticals.
• Fragmentation: Article 82 still allows national authorities to impose additional measures, risking a fragmented regulatory landscape across Member States.

WHAT IS THE EXPECTED LEGISLATIVE PROCESS?

The Digital Omnibus must proceed through the EU’s ordinary legislative procedure. Adoption is necessary before August 2026 to prevent the original high-risk requirements from taking effect without the supporting technical standards.

KEY TAKEAWAYS

The Digital Omnibus recalibrates the EU AI Act by addressing practical implementation hurdles. It offers a "standards-first" timeline and reduces paperwork for smaller entities and low-risk systems. While providing breathing room, businesses must remain ready for the "backstop" deadlines for high-risk compliance, with early legal assessment from experienced technology law firm and lawyers playing a critical role in navigating the evolving EU AI regulatory landscape.