The European Union has formally postponed the AI Act's main obligations for high-risk AI systems. The Council of the EU adopted the Digital Omnibus on AI on June 29, 2026, and the regulation was published in the Official Journal on July 24, 2026. Rules for stand-alone high-risk systems, which were due to apply from August 2, 2026, now apply from December 2, 2027. Other AI Act duties, including transparency rules for chatbots, were not postponed in the same way.
What changed
The European Parliament's Legislative Train page records the timeline. The Commission proposed the package on November 19, 2025, the Council and Parliament reached a provisional agreement on May 7, 2026, Parliament approved it on June 16, 2026 by 423 votes to 57 with 174 abstentions, and the Council adopted it on June 29. The main date changes are:
| Obligation | New application date |
|---|---|
| Stand-alone high-risk AI systems | December 2, 2027 |
| High-risk AI systems embedded in regulated products | August 2, 2028 |
| Marking of AI-generated content, for systems placed on the market before August 2, 2026 | December 2, 2026 |
The package also adds a prohibition on AI practices that generate non-consensual intimate or sexual content or child sexual abuse material, and it exempts AI-enabled machinery from direct application of the AI Act in favor of the sectoral Machinery Regulation approach.
Usercentrics' analysis notes that the regulation entered into force on July 27, 2026 and that the Article 50 transparency obligations were not amended. Duties such as informing people that they are interacting with an AI system therefore apply from August 2, 2026. Usercentrics also notes that generative AI systems placed on the market after August 2, 2026 do not get the grace period for content marking.
Why it matters for companies hiring IT firms
If you are commissioning AI features, such as CV screening, student assessment or customer chatbots, the delay changes your timeline but not the direction of travel. High-risk systems will still need risk management, technical documentation, human oversight and data quality controls; there is simply more time to build them in properly. Chatbots and generative features already face transparency duties. Contracts with development partners should state who is responsible for AI Act work, including documentation, user disclosures and labeling of generated content.
Useful questions to ask a vendor that builds AI features for you:
- Which AI models and providers does the feature use, and where is data processed?
- How are users told that they are dealing with an AI system or AI-generated content?
- What documentation will you hand over, and who updates it when the system changes?
- If the feature could fall into a high-risk category, what is the plan to meet the 2027 requirements?
Why it matters for IT firms
Software firms serving European clients should not treat the delay as a pause. Providers of generative AI systems need content marking in place by December 2, 2026 for systems already on the market, and immediately for new ones. Firms building AI into HR, education or finance products for the EU can use the extra time to build compliance steps into their delivery process, from data documentation to logging. Being able to explain that process clearly can help in proposals to EU clients.
What to do next
- List the AI features in your products and check whether any fall into high-risk categories.
- Confirm that chatbots and AI-generated content are disclosed and labeled as Article 50 requires.
- Plan for content-marking requirements by December 2, 2026.
- Add AI Act responsibilities to statements of work with development partners.
This article is a news summary, not legal advice. Looking for a partner with European experience? Browse IT companies in Europe or software development firms on Trusted IT Firms, or see all listed companies.
Frequently asked questions
Has the EU AI Act been delayed?
Only in part. Obligations for stand-alone high-risk AI systems now apply from December 2, 2027, and for high-risk AI embedded in regulated products from August 2, 2028.
Do chatbot transparency rules still apply from August 2026?
Yes. According to Usercentrics, the Article 50 transparency obligations were not amended and apply from August 2, 2026.
When must AI-generated content be marked?
Systems placed on the market before August 2, 2026 have until December 2, 2026. Systems placed on the market after that date get no grace period.
When was the Digital Omnibus on AI adopted?
The Council adopted it on June 29, 2026, and it was published in the Official Journal on July 24, 2026.
Sources
Add trusteditfirms.com as a preferred source on Google, or open this page in your AI assistant to use it as a source.
