What Does the EU AI Act Delay Mean for AI Used in Hiring?

Does the EU AI Act apply to a US company's HR systems?

It can. Jones Walker notes that the Act reaches providers and deployers established outside the EU "where the output produced by the system is used in the Union." A US company with employees or candidates in Europe, or a European team using a US-built hiring tool, should not assume the Act stops at the border. Separately, Jones Walker notes that most Article 50 transparency obligations, such as telling people when they are interacting with an AI system, applied from August 2, 2026, so those are already in force.

Why is this a CIO problem as much as an HR problem?

HR owns the decisions about people. The CIO is the person who can see which systems are making or shaping them. AI features now arrive inside tools the company already pays for, often switched on in an update, and teams adopt their own tools when the approved ones do not fit. Neither shows up in an HR policy. Both show up, eventually, in an integration log, a software inventory or an expense report.

The vendor relationship sits with the CIO too. Many AI features in hiring and performance tools are built by the vendor and switched on by the customer, which raises questions only a contract can answer: which features use AI, what the vendor considers its own obligations under the Act, and what documentation it will provide. Adding those questions to the next renewal costs nothing and produces the evidence counsel will want. Jones Walker also notes that high-risk AI embedded in regulated products under Annex I moves to August 2, 2028, which matters for some operations technology, though less for HR systems.

Average Robot calls the right response governed enablement: rules people can follow while they do their work, so the safe path is also the easy one. Applied here, it means the CIO and CHRO agree on a short, usable standard for any AI that touches people decisions, and make the approved route faster than the workaround. A rulebook that stops people working tends to push the same activity somewhere less visible.

This is the Direction link in the AI Profitability Gap, the chain Average Robot uses to trace AI spend through to return (how the chain works). Direction covers who decides what AI is pointed at. When AI is pointed at decisions about people, that choice needs a named owner on the HR side and a named owner on the technology side, and both need the same list of systems.

What should a CIO do with the extra time?

Each of these can start this quarter without new spend.

1. Build the inventory. List every system that uses AI to screen, rank, evaluate, schedule or monitor people, including features inside existing HR and collaboration tools.
2. Record where each system's output is used. Note whether any output reaches candidates or employees in the EU. That is the fact counsel will ask for first.
3. Note when each system was deployed and last changed. Orrick's point about substantial modification makes version history worth keeping now.
4. Assign an owner pair. For each system, name one HR owner and one technology owner, and give them the same review date.
5. Write one page of rules people can use. Say which AI uses in people decisions are approved, which need review and which are off limits, and make the approved route quick.

If you want a quick read on where your organization stands before that work begins, take the free AI Profit Readiness Assessment. It takes about two minutes, and your result appears on screen when you finish.