Glossary Updates12 new terms added to the glossaries · October 2, 2026, 22:44 CEST
AI TechDocKnowledge

Glossary · EU AI Act: foundations

Updated

Digital Omnibus on AI

Also known as: AI Omnibus, Digital Omnibus, Regulation (EU) 2026/1744

The Digital Omnibus on AI, Regulation (EU) 2026/1744, is the first amendment of the EU AI Act. Proposed by the European Commission in November 2025 and in force since July 27, 2026, it postpones the high-risk obligations to December 2, 2027 (Annex III) and August 2, 2028 (Annex I Section A), moves machinery to the sector route of Annex I Section B, adds new prohibited practices and simplifies obligations for smaller companies.

  • AI regulation
  • EU
  • Omnibus
  • Machinery

In one sentence

The Digital Omnibus on AI (Regulation (EU) 2026/1744) amends the AI Act: high-risk rules move to December 2027 and August 2028.

Example

A provider of an AI-based credit-scoring system, which planned for August 2026, now has until December 2, 2027 to meet the high-risk requirements.

What it changed

  • High-risk dates: Requirements for Annex III systems apply from December 2, 2027 instead of August 2, 2026, and for AI in products under Annex I Section A from August 2, 2028 instead of August 2, 2027 — fixed dates instead of the Commission's proposal to tie them to the availability of standards.
  • Machinery moves to Section B: The Machinery Regulation moves from Section A to Section B of Annex I (Recital 42). AI-related safety and conformity questions for machinery are assessed under the Machinery Regulation from January 20, 2027; AI-specific additions to its Annex III follow through Article 8 delegated acts by August 2, 2028, and a new Article 20(10) lets AI Act harmonized standards support conformity in the meantime. See the machinery AI timeline.
  • Safety component: The amended Article 6 narrows what counts as a safety component: integration into a regulated product is not enough, and pure assistance, optimization or convenience functions are excluded unless their failure would endanger health or safety.
  • New prohibitions: AI generating non-consensual intimate imagery and child sexual abuse material, prohibited from December 2, 2026.
  • Transparency: Article 50 still applies from August 2, 2026; only the marking duty for systems already on the market gets a grace period until December 2, 2026.
  • Simplification: Relief on documentation and quality management for smaller companies, a replaced Article 4 on AI literacy (providers and deployers take measures to support it instead of ensuring a sufficient level) and wider supervision by the AI Office.

How it applies

  • Technical documentation: The postponement is time to prepare, not an exemption. Harmonized standards for high-risk AI are expected in the meantime, and documentation built to them now will be ready when the obligations apply.
  • Machinery: August 2, 2028 is not the first date for AI in machinery. The machinery-law assessment of AI functions that affect safety or control begins on January 20, 2027.
  • Citing the law: Refer to the AI Act “as amended by Regulation (EU) 2026/1744”, or use the consolidated text on EUR-Lex, to avoid quoting outdated dates.

Compared with the USA, Canada and China

The Omnibus is one of several EU omnibus packages to reduce regulatory burden, in a year when the USA sought to preempt state AI laws and Canada chose not to revive AIDA. It delays and simplifies the AI Act but keeps its structure; the EU remains the only one of the four with a comprehensive AI law.

Read more on AI TechDoc Press

The in-depth analysis behind this entry, in the AI TechDoc Press newsletter. Subscribe for free