August 2026: the month AI hiring stopped being unregulated
The EU AI Act treats recruitment AI as high-risk. New York audits it annually. Colorado wants reasonable care. Here is the checklist procurement will hand you.
Three regimes now shape how AI can be used to hire. The EU AI Act classifies recruitment and employee-selection systems as high-risk, with obligations phasing in and significant penalties attached. New York City's Local Law 144 requires annual bias audits of automated employment decision tools, with results published and candidates notified. Colorado's AI Act, effective June 2026, requires developers and deployers to exercise reasonable care against algorithmic discrimination. The EEOC continues to signal enforcement interest.
None of this bans AI in hiring. All of it assumes you can explain what your system did and why.
The three questions before any feature question
What data trained this. Who can see the result. Can we reproduce this decision in a year. Every enterprise procurement conversation starts there, and each answer is an architecture property rather than a policy statement: role-based permissions and read/write scoping through Access Control, defined retention windows, and audit logs plus batch reporting that record every score, every override and every reviewer.
The checklist
Inventory every automated tool touching a hiring decision, including the ones a single team bought. Record what each one scores and what it explicitly refuses to infer. Confirm a human owns every negative decision, and that overrides are logged. Keep candidate notice and, where required, consent — plus a working deletion path. Commission the bias audit where a jurisdiction requires one, and diarise its renewal. Then keep an evidence pack: model documentation, audit results, retention policy, access matrix.
The trust gap is the business case
Gartner found only 26% of candidates trust AI to evaluate them fairly, while Pew Research reports a majority of Americans would not apply to an employer using AI to make hiring decisions. Whatever the regulation requires, the candidate market is already pricing in opacity. Publishing what you score — and what you refuse to score — is the cheapest trust you can buy.
This is general guidance, not legal advice; requirements differ by jurisdiction and change quickly. HireSuite.ai supports enterprise deployments with SSO, role-based access, approval workflows, audit logging and white-label candidate experiences — see Solutions or talk to an expert.
Amina Yusuf
Counsel, Responsible AI
Amina advises on responsible AI and hiring compliance at HireSuite.ai.
Comments (0)
Related reading
All articles
Rank, don't reject: the only safe way to point AI at a CV pile
Aug 6, 2026 · 5 min read
Nine in ten recruiters have suspected a candidate of faking it. Now what?
Jul 29, 2026 · 8 min readSeven things a voice can tell you about an interview — and four it can't
Jul 8, 2026 · 7 min readTwo Products, One Platform
Hiring teams run AI interviews, screening and analytics. Job seekers build AI-enhanced CVs, apply across thousands of live jobs and practise interviews.
12
Articles
50K+
Readers
4K+
Hires Studied