Skip to main content

What the EU Digital Omnibus Changes in the AI Act

21 NOVEMBER 2026·3 MIN READ·4 SOURCES

The EU Digital Omnibus on AI entered into force on 27 July 2026 and amends the EU AI Act, Regulation (EU) 2024/1689. High-risk AI system obligations will apply later than originally planned, but other AI Act requirements are already in force.

Key takeaways · 4

  • 01

    The Omnibus entered into force on 27 July 2026 and amends the EU AI Act.

  • 02

    Prohibited-practices rules can carry fines of up to €35 million or 7% of worldwide annual turnover, whichever is higher.

  • 03

    Providers of generative AI systems placed on the market before 2 August 2026 have until 2 December 2026 to meet Article 50(2)’s marking obligation.

  • 04

    Annex III systems, which include areas such as recruitment, creditworthiness and education, apply from 2 December 2027.

A change to timing and obligations

The Digital Omnibus on AI, Regulation (EU) 2026/1744, was adopted by the European Parliament and Council on 8 July 2026 and entered into force on 27 July.[1][3] It amends the AI Act, Regulation (EU) 2024/1689.[1] High-risk AI system obligations will apply later than originally planned.[1] But Levitina says the measure is not just a postponement: parts of the AI Act already apply, and the Omnibus also adds prohibitions and enforcement tools.[1] The Commission says the package is intended to ease compliance for smaller businesses, extend timelines, expand testing opportunities and provide greater legal clarity.[4]

Rules already in effect

The AI Act’s first provisions have applied since 2 February 2025.[1] Prohibited practices include manipulative or deceptive techniques, social scoring, untargeted facial-image scraping, and workplace and education emotion recognition, subject to limited exceptions.[1] Infringements of those rules may lead to fines of up to €35 million or 7% of worldwide annual turnover, whichever is higher.[1] Providers of general-purpose AI models have been subject to Chapter V since 2 August 2025; their obligations include maintaining technical documentation, putting a copyright policy in place and publishing a summary of training content.[1] The Commission’s power to fine those providers has applied since 2 August 2026.[1]

Transparency deadlines to track

The AI Act’s Article 50 transparency obligations have applied since 2 August 2026.[1] Users must be told when they interact with an AI system, synthetic content must be marked in a machine-readable format, and deepfakes must be disclosed.[1] There is a specific later deadline for one marking requirement: providers of generative AI systems placed on the market before 2 August 2026 have until 2 December 2026 to comply with Article 50(2)’s marking obligation.[1] Teams should verify which products and content workflows fall within that provision rather than assuming all requirements share one deadline.

High-risk systems and support measures

The systems listed in Annex III apply from 2 December 2027.[1] Levitina identifies areas covered by Annex III including recruitment, creditworthiness, insurance risk assessment and pricing, education and biometrics.[1] The Commission says some measures previously limited to small and medium-sized enterprises are extended to small mid-cap companies, and that the Omnibus expands access to regulatory sandboxes and introduces an EU-level sandbox.[4] The regulation cites delays in preparing standards and establishing national governance and conformity-assessment frameworks as reasons the compliance burden was heavier than expected.[3] It says the measures are intended to operate without reducing protections for health, safety and fundamental rights.[3]

For professionals developing, procuring or using AI in the EU, a delayed high-risk timeline does not remove the need to manage obligations that already apply. Use the relevant dates and system categories to prioritize a compliance review, and confirm whether your products or workflows fall under the specific provisions described.

Why it matters
For teams

AI training for your team

kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.

How this developed

  1. 21 November 2026

    What the EU Digital Omnibus Changes in the AI Act

Sources

AI fluency, one session a day, built for your work.