EU AI Act High-Risk Deadline Moves to December 2027 — But Not Everything Waits
Not legal advice. This article discusses legal and regulatory developments and is generated with AI assistance under human editorial review. It is provided for informational purposes only and does not constitute legal advice. Laws and regulations vary by jurisdiction and are subject to change. You should consult your own qualified legal counsel before taking or refraining from any action based on the content of this article. Read our AI Transparency statement.
The News
The European Parliament and Council of Ministers have finalised legislation delaying the EU AI Act's high-risk AI system obligations from 2 August 2026 to 2 December 2027 for stand-alone systems, with product-embedded systems following on 2 August 2028. The delay, part of the AI Omnibus simplification package, applies to Annex III systems including AI used in recruitment, screening, candidate evaluation, and employment decisions. The text is awaiting formal publication in the EU's Official Journal. Transparency obligations for AI-generated content are separately delayed to 2 December 2026, while prohibitions on certain AI practices have been in effect since February 2025. The Act has extraterritorial reach: employers outside the EU using AI to screen EU-based candidates remain in scope regardless of where they are headquartered.
Why It Matters
The delay is a reprieve, not an exemption. The compliance obligations themselves — risk assessments, bias testing, human oversight, technical documentation, transparency disclosures — have not changed, only the timeline. TA leaders who were racing to meet the August deadline now have 16 additional months, but the risk is that organisations treat the delay as a reason to deprioritise AI governance entirely. The employers best positioned for December 2027 will be those who used the extra time to inventory their AI tools, map decision points, and build oversight processes — not those who filed the project under "next year's problem."
Taking Action
All action is situated in your corporate context. These are starting points, not prescriptions — choose what fits.
Connect
Who to talk with in your enterprise
Your data protection officer and HRIS lead. Ask: do we have an inventory of every AI system that touches recruitment or employment decisions? Which of those systems screen, rank, or score candidates? The DPO needs this for GDPR alignment; the HRIS lead needs it for vendor management and contract renegotiation.
Inspect
What to take a closer look at
Pull your vendor contracts for any AI-powered hiring tool. Check whether they include provisions for bias testing, transparency, change notifications, and audit support. If they don't, that's your first conversation — and you now have 16 months to have it.
Reflect
What to think about more deeply
If your AI governance plan was built around an August deadline that just moved 16 months, what exactly were you going to do in those 16 months that you couldn't do in the time you already had?