
EU high-risk AI database pushed to mid-to-late 2027: a planning headache for builders
Published by AINave Editorial • Reviewed by Ramit
The European Commission has pushed back the launch of the central EU database for high-risk AI systems under the AI Act to the second half of 2027, creating uncertainty for compliance planning. The database is a key requirement for registering high-risk AI systems before deployment, and the delay means developers must navigate shifting deadlines.
What happened
The Commission's AI Act service desk confirmed in a message seen by Euractiv that the database is not yet open and operational. The AI Office inside the European Commission is responsible for building the register. Euractiv understands that the database will be officially launched in the third quarter of 2027.
The service desk stated that the obligation to register high-risk AI systems applies from December 2027, aligning with the Digital Omnibus simplification package's adjusted timeline. However, the Omnibus did not move a separate standalone article setting out the duty to register, which could imply an August 2, 2026 deadline. The Commission did not respond to requests for clarification.
Why AI builders should care
If your product uses AI in sectors like education, law enforcement, or workforce management, it may qualify as high-risk under the AI Act. The database timeline directly affects when you must register those systems. With the database not ready until late 2027, you cannot register earlier even if the legal obligation starts sooner. This creates a gap between legal requirements and operational capability.
The delay also means compliance roadmaps built around earlier deadlines (like June 2027 or August 2026) need revision. Builders should treat the current timeline as provisional and monitor for official updates from the AI Office.
Practical implications
For now, the practical registration deadline is December 2027, per the service desk. But the ambiguity around the standalone article means you should prepare for a possible August 2026 obligation if the database becomes available earlier. Daniel Leufer of Access Now called the database a key element of the law's architecture that must be ready when obligations kick in.
Start internal readiness planning now: classify your AI systems, document compliance measures, and track EU guidance. The delay gives you more time, but the underlying requirements are not changing.
Caveats
All dates cited by the service desk are subject to change by EU authorities. The Digital Omnibus revisions are still being implemented, and further adjustments are possible. The standalone article ambiguity has not been resolved by the Commission. Builders should not rely solely on the December 2027 date without monitoring official publications.
FAQs
Sources
- EU’s high-risk AI database pushed back to mid-to-late 2027
- EU to delay high-risk AI rules after industry pressure – Business...
- Brussels Is Rewriting the Clock on High-Risk AI... - Captain Compliance
- He Risked Everything To Warn You: No One Is Ready For... - YouTube
- Why the EU's AI Act Delay Is Actually a Lesson in How to Govern Tech
- The EU Just Blinked on Its Own AI Rules. Use the Window, Do Not...
- EU AI Act Enforcement Is Here: Chatbot Rules Live, High-Risk AI Delay Now Binding Law
- Workplace AI — how employers should prepare for the new EU AI Act deadline
- EU delays high-risk AI rules to 2027 and 2028 — AI-Brainer
- INTERVIEW: EU must not repeat Ireland’s mistakes on data... | Euractiv
- Ethics as an Engine, Not a Brake: Why Europe’s "Simplification..."
- AI2027: Is this how AI might destroy humanity? - YouTube
- A Guide to AI 2027 — LessWrong
- An Ex-OpenAI researcher predicts AGI by 2027—here’s the wild...
- AI Isn't Helping Plan the Break-In Anymore. It Is the Break-In.
- EU Commission drafts guidelines on classifying high-risk systems under the AI Act






















