The high-risk deadlines moved. The basics did not.

What happened

The AI Act (Regulation (EU) 2024/1689) entered into force on 1 August 2024. The prohibited practices and the AI-literacy provision have applied since 2 February 2025, and the obligations for general-purpose AI models since 2 August 2025. According to the European Commission’s AI Act Service Desk, most remaining rules have applied since 2 August 2026, including the transparency rules of Article 50, and enforcement by national and EU authorities has started.

The Commission proposed a Digital Omnibus on AI on 19 November 2025. Parliament and Council reached a provisional agreement on 7 May 2026, Parliament approved it on 16 June and the Council adopted it on 29 June. The amending regulation was published in the Official Journal on 24 July 2026 as Regulation (EU) 2026/1744 and, according to the Commission, entered into force on 27 July 2026. It is law, not a proposal.

What the amendment changed

• High-risk rules postponed. Rules for high-risk systems in the areas of Annex III, such as employment and critical infrastructure, apply from 2 December 2027; for high-risk AI built into products covered by Annex I, from 2 August 2028.
• AI literacy reworded. Providers and deployers must take measures to support the AI literacy of their staff and others using AI on their behalf. The text states that this does not require guaranteeing a specific level for any individual; the Commission and Member States must support companies, in particular SMEs.
• New dates in December 2026. A ban on AI systems that generate non-consensual sexual content or child sexual abuse material, and a transitional deadline for marking AI-generated content from systems already on the market before 2 August 2026.
• Other changes. Simplifications for SMEs extended to small mid-cap companies, wider access to regulatory sandboxes and a clarified relationship with the Machinery Regulation.

This summary reflects the consolidated text published by the Commission’s AI Act Service Desk as of 27 July 2026. Guidance and practical examples continue to be published; check the current text for your own case. This article is not legal advice.

What it shows for a first AI process

The postponement gives companies with high-risk use cases more time. It does not change the starting point for everyone else. Whether a use falls in a high-risk category depends on what the system is used for, not on the technology, so each use case needs its own assessment. And the duty to support AI literacy already applies: people who use AI at work should understand what the system does, where it can be wrong and when to check.

A controlled first process makes those questions concrete. Choose one recurring task with known documents, such as reading incoming requests or searching technical manuals. Name the users and the person responsible. Record which system is used, for which purpose and with which data. Make sure every output can be traced to its source and that a person accepts, corrects or rejects it. If customers or the public interact with the system or receive generated content, check which transparency rules apply.

Where Meridian fits

Meridian AI is assistive: it reads documents, answers with references to the source, summarises and prepares drafts. It is designed to run in your own environment, with people reviewing every result before it is used. That makes a process easier to explain, check and stop. It does not make an organisation compliant with the AI Act; that remains an assessment for each use case, with your own legal and compliance advisers. This article does not suggest that any company named in the sources uses Meridian.

Start with one AI process you can explain, check and stop.