Fairness · NYC LL 144 · EU AI Act
Fairness isn't a polish pass.
Chiron's technical screening is designed against NYC Local Law 144 and the EU AI Act from the foundations up. Every recommendation is reviewed by a human; every dimension shows its evidence.
The frameworks we hold ourselves to
WCAG 2.1 AA
Accessible by design
Built for accessibility from the start: sufficient contrast, full keyboard operability including the workspace, reduced-motion support, and never color as the only signal — considered every release.
◆Color-blind-safe result bands
◆Reduced-motion fallbacks
◆Full keyboard operability
◆Visible focus throughout
NYC Local Law 144
Built to be bias-audited
Chiron is designed against NYC LL 144's four-fifths rule so its outcomes can be independently audited for disparate impact. Every scored dimension is auditable, and every recommendation is reviewed by a human before it advances a candidate.
◆Designed against the four-fifths rule
◆Every dimension auditable
◆Human-reviewed recommendations
◆No hidden or automated verdicts
EU AI Act · high-risk
Article 14 oversight, by design
Hiring is high-risk under the Act. Chiron meets Article 14 (human oversight) not by toggle but by structure — the reviewer's decision is always required to advance a candidate.
◆No auto-advance, ever
◆Recommendation, never a verdict
◆A human always makes the call
Evidence
No score stands alone
Every dimension expands to the quote, diff, or moment that produced it. Reviewers always have the grounds behind a recommendation — there is no black-box number.
◆Cited evidence per dimension
◆Full session available to review
◆Reviewers see the why
◆Candidates can request their data