As of September 2026. This article reflects our understanding and is not legal advice. Whether and how the rules affect your business is something to clarify with a lawyer or your data protection officer.

What has been adopted

With the so-called Digital Omnibus, the EU has revised several digital laws, including the AI Act. The amending regulation is numbered (EU) 2026/1744. It was published on 24 July 2026 and has been in force since 27 July 2026.

For most businesses, three points matter.

AI literacy: reworded, not abolished

Since February 2025, Article 4 of the AI Act has required companies that use AI to do something for their employees’ AI literacy. Until now they were to ensure that their staff have a sufficient level of AI literacy.

As we understand it, companies must now take measures to promote their employees’ AI literacy. They no longer have to guarantee a particular level for each individual. Experts do not yet agree on every point of interpretation.

In practice little changes for a business. Whoever approves AI tools also explains them, writes down rules and names a contact person. What that involves is described in the guide AI literacy under the AI Act.

High-risk AI: later

The strictest duties in the regulation apply to so-called high-risk AI, for example systems that help decide on job applicants, loans or access to services. These duties were due to apply from August 2026. The Omnibus has postponed them:

  • For most high-risk systems (Annex III of the regulation) to 2 December 2027.
  • For AI in certain regulated products (Annex I), such as machinery or medical devices, to 2 August 2028.

For a business that uses AI for writing, summarising or searching, this usually makes no difference. You should pay attention if AI is to help decide about people. Such plans need a legal review beforehand, even if the duties only apply later.

Transparency: already in force

The transparency duties under Article 50 have applied since 2 August 2026. Put simply: anyone talking to an AI system should be able to tell, and certain AI-generated content should be recognisable as such. For systems already on the market before then, there is a transition period until 2 December 2026 for marking generated content (Article 50(2)).

For businesses this matters above all when customers come into contact with AI, for example through a chatbot on the website. What to watch out for is described in the guide A chatbot on your website.

What you should do now

  1. Keep track. Which AI tools are used in the business, by whom and for what?
  2. Do not put off rules and briefings. The AI literacy duty still applies. How to organise the use of ChatGPT and Copilot is shown in the guide ChatGPT and Copilot at work.
  3. Disclose where customers talk to AI. A sentence at the start of a chat is usually enough.
  4. Take care with decisions about people. Selecting applicants, assessing employees: these need a legal review before use.

The postponement gives providers of high-risk systems some breathing space. For most businesses the task stays the same: use AI where it helps and make sure employees know how. Where that could be in your business is what we work out in our AI consulting.