REGULATORY INTELLIGENCE · ODA3 INSIGHTS

EU AI Act Deadlines Shift: Omnibus Package Extends High-Risk Compliance to 2027–2028

The European Parliament has adopted its negotiating position on the Digital Omnibus package, which proposes targeted amendments to the AI Act deadlines ..

Editorial header for EU AI Act Deadlines Shift: Omnibus Package Extends High-Risk Compliance to 2027–2028
CATEGORYRegulatory Intelligence
EVIDENCE BASISSource publication
PUBLISHEDApril 21, 2026
READING TIME4 min

Article

Target Audience: Compliance Officers, AI Governance Leads, Legal Counsel
Category: Regulatory Update / Compliance

Executive Summary:
The European Parliament has adopted its negotiating position on the Digital Omnibus package, which proposes targeted amendments to the AI Act deadlines . High-risk Annex III systems now have until December 2, 2027, and sectoral product safety systems until August 2, 2028. However, transparency obligations for generative AI begin November 2, 2026, and the the amended EU AI Act timetable enforcement wave is still active. Critically, a potential loophole exists: high-risk systems placed on the market before the deadlines may fall outside obligations unless substantially modified.

Evidence Tier: Secondary Verified (Stephenson Harwood law firm analysis of European Parliament action)

What’s Changing

The European Parliament has adopted its negotiating position on the European Commission’s “Digital Omnibus” package, which proposes to amend the AI Act .

The text is not yet law. Trilogue discussions (European Parliament, Council of the EU, European Commission) begin April 28, 2026 .

But the direction is clear: deadlines are extending for some obligations, but NOT all.

Updated Deadlines

Obligation CategoryNew Proposed DeadlineOriginal Landscape
GPAI (general-purpose AI) transparency and copyrightStill the amended EU AI Act timetableNo change
Generative AI content transparency (machine-readable detection)November 2, 2026New date
High-risk Annex III systems (biometrics, critical infrastructure, education, employment, law enforcement, etc.)December 2, 2027Extended
EU sectoral product safety regimes (medical devices, radio equipment, toy safety)August 2, 2028Extended

The Critical Loophole (and Risk)

The law firm Stephenson Harwood highlights a potential key issue in their analysis of the Omnibus package :

“A potential key issue arises from the combination of delayed application dates and non-retrospective operation of the AI Act. High-risk AI systems placed on the market before the relevant deadlines would generally fall outside the high‑risk obligations, unless they are subsequently substantially modified. This could incentivise providers to deploy high‑risk systems ahead of late 2027, with the effect that such systems may remain in use without being brought within the full scope of the AI Act’s high‑risk regime.”

Translation: There is now a window to rush high-risk AI systems to market before December 2027 and potentially avoid compliance obligations for the life of those systems (absent major modifications).

This is not a compliance strategy. This is a risk creation strategy.

New Prohibition: “Nudifier” Apps

The European Parliament has also backed an explicit prohibition on “nudifier” AI systems that generate or manipulate realistic sexually explicit or intimate images of identifiable persons without consent, with narrow exceptions for effective technical safeguards that prevent misuse .

What Hasn’t Changed

The the amended EU AI Act timetable enforcement wave is still active. Organizations that have been preparing for:

  • GPAI provider obligations
  • Copyright disclosures
  • Fundamental rights impact assessments for high-risk systems (even if enforcement is delayed)

…must continue on track. Delaying compliance because deadlines shifted is a governance failure.

Actionable Compliance Roadmap

TimelineAction
Immediate – the amended EU AI Act timetableGPAI transparency and copyright obligations. Generative AI content must be machine-readable and detectable as artificially generated by November 2, 2026 .
August – November 2026Implement content provenance and detection systems.
2027Complete high-risk Annex III compliance assessments. Systems placed on market in 2027 will be subject to delayed deadlines but should be built compliant now.
2028Sectoral product compliance for medical devices, radio equipment, etc.

The Bottom Line for CISOs and Compliance Officers

Do not treat deadline extensions as permission to slow down. The compliance burden is not reduced—only the enforcement clock moved. Organizations that build compliant now will have market advantage over those who rush non-compliant systems to market before the deadlines.

Second-order risk: If you acquire or integrate a system placed on market before December 2027 that was rushed to avoid compliance, you inherit that risk when that system is “substantially modified” or when your use case triggers obligations independently.

Framework context

This article supports operational interpretation across the GAISSF Ecosystem. Use GAISSF for governance and assurance context, UAIF for incident classification, and AI-IRF for incident-response architecture. These links describe relationships; they do not assert certification, regulatory approval, or legal compliance.

Tags

AI SecurityRegulatory IntelligenceODA3 InsightsEU AI ActCompliance

Continue reading