EU AI Act update, 9 May 2026: current law remains the baseline. The Digital Omnibus provisional agreement would move many high-risk AI obligations to 2 Dec 2027 and product-integrated high-risk AI rules to 2 Aug 2028 if formally adopted. Track status EU AI Act update: current law remains the baseline. Digital Omnibus dates apply only if formally adopted. Track status

Free EU AI Act Digital Omnibus planning tool

EU AI Act Omnibus Deadline Router

The 7 May 2026 Digital Omnibus agreement did not create one simple new AI Act deadline. It created several provisional planning tracks. Use this router to identify the likely track for one AI system and the evidence work you should not pause.

Run the deadline router Read the Omnibus status note

Educational planning tool only. Not legal advice. Not a compliance guarantee. Current published law remains Regulation (EU) 2024/1689 until amended text is formally adopted and published.

EU AI Act Omnibus deadline router showing 2026, 2027 and 2028 planning tracks for AI systems
A deadline router is useful because the Omnibus creates different planning tracks, not a single pause button.

Quick answer

Ask a better question than “Was the AI Act delayed?”

The operational question is: which obligation track applies to this AI system, and what evidence should still be built now? The router below separates the current-law baseline from provisional planning tracks for stand-alone high-risk systems, product-integrated high-risk systems, transparency triggers, and prohibited-practice risk.

Use it for triage before board updates, vendor reviews, product release decisions, and AI inventory cleanup. Do not use it as a legal classification engine.

Router tool

Find the likely planning track

Answer for one AI system or planned AI feature. The result gives a planning track, records to keep, and the next EU AI Compass tool to open.

Privacy note: this static router runs in your browser and does not require login. Do not enter confidential, personal, or privileged information.
1. What is your main role for this AI system?
2. Which description fits the AI system?
3. Any immediate risk flags?
Compliance team mapping EU AI Act current-law baseline and Digital Omnibus planning tracks into an evidence file
The useful output is not a date. It is a dated planning note plus evidence actions.

Decision table

The Omnibus creates tracks, not a general pause

QuestionLikely planning track to reviewWork not to pause
Possible stand-alone high-risk AI?2 December 2027 if the provisional agreement is formally adopted and published.Inventory, role decision, intended purpose, risk route, oversight owner, vendor evidence.
AI embedded into a regulated product?2 August 2028 if the provisional agreement is formally adopted and published.Product-safety overlap, technical file, conformity route, sectoral-law map, change log.
User-facing or synthetic content system?Article 50 and transparency solution review. Do not assume this moved with high-risk dates.Notice trigger review, content marking decision, user-facing wording, evidence record.
Nudification, CSAM, or intimate-content risk?Prohibited-practice watch item. Treat as urgent governance risk, not a later-deadline issue.Block/disable review, vendor questions, misuse controls, escalation path, source note.

Evidence actions

What to document even if a later date becomes law

AI inventory

System name, owner, purpose, vendor, data category, user group, and change status.

Role decision

Provider, deployer, product manufacturer, importer, distributor, or unclear owner path.

Trigger review

Annex III, Article 50, prohibited-practice, product integration, and vendor trigger notes.

Source log

Date reviewed, source used, current-law baseline, provisional-agreement assumption, and decision owner.

Digital Omnibus cluster

Use the cluster in the right order.

Start with the router, then open the guide that matches the planning track: standards readiness, product-integrated AI, or SME and small-mid-cap relief.

TOOL · 3 MIN · CURRENT PAGE

Omnibus Deadline Router

Separate current-law dates from 2026, 2027, and 2028 planning tracks for one AI system.

 GUIDE · 5 MIN

Standards Readiness After the Omnibus

Build evidence structures that can absorb harmonised standards and common specifications when they arrive.

 GUIDE · 5 MIN

Product-Integrated AI Systems and 2028

Connect AI Act planning with the product safety file, supplier evidence, and sectoral-law overlap.

 GUIDE · 5 MIN

Small Mid-Cap Relief After the Omnibus

Treat SME and small-mid-cap relief as a planning issue, not an exemption from records.

Frequently asked questions

Use these answers to separate current law, provisional-agreement planning, and evidence actions before relying on a deadline assumption.

Not as a blanket rule. The 7 May 2026 Digital Omnibus agreement is provisional and still requires formal adoption and publication. The router separates current-law baseline work from if-adopted planning tracks, including 2 December 2027 for many stand-alone high-risk systems and 2 August 2028 for product-integrated high-risk systems.
No. The Council describes the 7 May 2026 outcome as a provisional agreement that must still be endorsed, legally and linguistically revised, and formally adopted. EU AI Compass treats the current published Regulation (EU) 2024/1689 as the baseline until amended text is formally adopted and published.
The provisional agreement points to stand-alone high-risk AI systems in areas such as biometrics, critical infrastructure, education, employment, migration, asylum, and border control. Treat this as a planning track, not a final legal classification. Keep the AI inventory, role decision, intended purpose, and evidence owner current.
The provisional agreement points to high-risk AI systems integrated into regulated products, including examples such as lifts or toys, with sectoral-law overlap also discussed for areas such as machinery, medical devices, toys, lifts, and watercraft. Product teams should map sectoral safety law and AI Act evidence together.
Yes. Article 50 transparency issues need separate review and should not be treated as automatically delayed by high-risk system timing. Chatbots, deepfakes, AI-generated content, biometric categorisation, and emotion-recognition use cases still need a trigger review, notice decision, and evidence record.
No. Deployers should use any extra time to improve the evidence file, not pause it. Useful records include AI inventory fields, vendor evidence, user instructions, oversight notes, Article 50 trigger decisions, training records, and dated source-review notes showing which legal baseline was used.
The router result is an operational planning output, not a legal opinion. It helps a team identify the likely track to review, the evidence to keep, and the related EU AI Compass tools to open. Final classification and legal reliance should be reviewed with qualified counsel or the relevant authority.
No server submission is used by this static page. The router runs in the browser and displays the result on the page. Do not enter personal data, confidential business information, or legally privileged material into public webpages.

Source basis

Last reviewed: 19 May 2026.

Source basis: Regulation (EU) 2024/1689, European Commission AI Act page and 7 May 2026 Digital Omnibus press release, Council 7 May 2026 provisional-agreement press release updated 18 May 2026.

Status caveat: The 7 May 2026 Digital Omnibus outcome is treated as a provisional agreement and planning track until formal adoption and publication are complete.

Use limitation: This page is educational and operational. It does not provide legal advice, determine legal status, confirm exemption, or prove compliance.