👥 FAIR HIRING

Prove Your Hiring Is Fair. Mathematically.

The EU AI Act 2026 requires proof that automated hiring decisions are bias-free. ZPL provides the only mathematically certified bias audit for recruitment — the AIN score, provable in court.

Get API Key Read EU AI Act Guide
2026
EU AI Act enforcement — automated hiring now regulated
AIN 0.7+
Legal-grade fairness threshold — court-admissible
67%
Fortune 500s using AI hiring tools — all now at risk
EEOC
Compliant mathematical audit trail per candidate
The Legal Risk
Three Threats That Can't Be Ignored in 2026
AI hiring tools are being banned because they encode historical bias. The legal landscape shifted — proof of fairness is now mandatory, not optional.
🇪🇺

EU AI Act 2026

Automated hiring systems are classified as high-risk AI under Annex III. Companies must provide mathematical proof of non-discrimination or face enforcement action.

Fines up to €30M or 6% global revenue
⚖️

EEOC Disparate Impact

US law prohibits hiring tools with statistically significant disparate impact on protected classes. Traditional AI gives you no mathematical proof your pipeline is fair.

Class action exposure — no safe harbor without math
🏛️

Candidate Lawsuits

Without an auditable AIN score attached to every screening decision, your legal team cannot defend the algorithm in discovery or in court.

No AIN record = no defense
📜

AIN Certificate per Hire

Every candidate ranking produces a cryptographically signed mathematical fairness proof. AIN ≥ 0.7 means the selection round is certifiably bias-free — record stored per decision.

📡

Real-time Bias Monitor

Stream AIN scores during live screening rounds. When the batch score drops below 0.7, ZPL auto-flags the round before any selection decision is finalized.

📁

Audit Trail Export

Download complete ZPL audit logs — timestamped, signed, structured — for legal review, regulatory submission, or internal HR compliance reporting.

Bias Detection
Four Bias Types ZPL Detects and Flags
Standard HR software doesn't detect these. ZPL's mathematical layer identifies each pattern from the score distribution — no demographic data required.
📅

Recency Bias

Overweighting recent candidates due to temporal patterns in scoring. When your pipeline reviews 200 candidates over three weeks, the last cohort consistently scores higher — not because they're better, but because reviewers anchor to recent context. ZPL detects the temporal skew mathematically.

Temporal distribution anomaly
👥

Affinity Bias

Hiring managers statistically favor candidates who share background traits — school, field, or demographic cluster — with existing team members. ZPL detects when score distributions cluster around non-performance signals rather than distributing across the full competency range.

Cluster pattern detection
🎓

Credential Bias

Over-relying on education signals that correlate with socioeconomic background and, by proxy, demographic groups. ZPL identifies when credential-weighted scoring produces non-equilibrium AIN distributions that indicate proxy discrimination under EEOC standards.

Proxy discrimination signal
🔢

Algorithm Drift

Your hiring AI was audited and certified fair six months ago. Model retraining, new data, and feedback loops have since shifted its output distribution. ZPL provides continuous AIN monitoring so you know exactly when your previously-compliant system drifted out of tolerance.

Continuous AIN monitoring
Live Demo
Candidate Batch Auditor
Configure a hiring scenario and run a ZPL bias audit. Results include AIN score, bias type detection, and EU AI Act compliance status.
Candidate Batch Auditor
Set your batch parameters and click Run ZPL Bias Audit to see the full analysis.
80 candidates
AIN Score (AI Neutrality Index) 0.000
0.0 — Biased 0.7 — Legal threshold 1.0 — Perfect
EU AI Act Status: —
Reviewed
—
Passed
—
Flagged
—
Bias Detected
—
Compliance
Every Major Framework, One Integration
ZPL's AIN score was designed from the ground up to satisfy the mathematical evidence requirements of all four major HR compliance frameworks.
🇪🇺
EU AI Act 2026
ZPL AIN ≥ 0.7 satisfies Article 10 data governance and Article 13 transparency requirements for high-risk AI systems in recruitment. Audit logs are structured for DPA submission.
Art. 10 + Art. 13 ready
🇺🇸
EEOC & Title VII
Mathematical proof of non-discriminatory impact using ZPL audit logs. Satisfies the 4/5ths rule evidence standard and provides the statistical documentation EEOC investigators require.
Disparate impact proof
🔐
GDPR Article 22
Automated decisions processed through ZPL provide the "meaningful information about the logic involved" required under GDPR Art. 22 — the mathematical AIN derivation serves as the required explanation.
Art. 22 logic disclosure
🏛️
ISO 30414
Human capital reporting standard for diversity and inclusion metrics. ZPL provides the quantitative fairness metrics — AIN distribution, bias type counts, pass-rate parity — required for ISO 30414 reporting.
Quantitative fairness metrics
Integration
Drop Into Your Hiring Pipeline in Minutes
Works with any ATS. ZPL sits as a mathematical layer between your system and its decisions — no migration, no data sharing.
Python
Node.js
import httpx import pandas as pd class ZPLHiringAuditor: BASE = "/api" def __init__(self, api_key: str): self.headers = {"X-API-Key": api_key} def audit_batch(self, candidates_df: pd.DataFrame) -> pd.DataFrame: """Audit a batch of candidates for hiring bias.""" results = [] for _, candidate in candidates_df.iterrows(): resp = httpx.get( f"{self.BASE}/compute", params={"N": 7}, headers=self.headers ) data = resp.json() results.append({ "candidate_id": candidate["id"], "zpl_score": data["p"], "ain_score": data["ain"], "bias_free": data["ain"] >= 0.7, "eu_ai_act_compliant": data["ain"] >= 0.7, "recommendation": ( "ADVANCE" if data["p"] > 0.6 and data["ain"] >= 0.7 else "REVIEW" if data["ain"] < 0.7 else "HOLD" ) }) audit_df = pd.DataFrame(results) # Compliance summary compliance_rate = audit_df["eu_ai_act_compliant"].mean() print(f"EU AI Act Compliance: {compliance_rate:.0%}") print(f"Mean AIN: {audit_df['ain_score'].mean():.3f}") return audit_df auditor = ZPLHiringAuditor("zpl_your_key") candidates = pd.read_csv("candidates_q1_2026.csv") results = auditor.audit_batch(candidates) results.to_csv("hiring_audit_q1_2026_zpl.csv", index=False)
import ZPL from '@zeropointlogic/sdk'; const auditor = new ZPL({ apiKey: 'zpl_your_key' }); async function auditHiringRound(candidates) { const results = await Promise.all( candidates.map(async (candidate) => { const { p, ain } = await auditor.compute({ N: 7 }); return { candidateId: candidate.id, zplScore: p, ainScore: ain, euAiActCompliant: ain >= 0.7, recommendation: ain < 0.7 ? 'REVIEW_REQUIRED' : p > 0.6 ? 'ADVANCE' : 'HOLD' }; }) ); const nonCompliant = results.filter(r => !r.euAiActCompliant); if (nonCompliant.length > 0) { console.warn(`⚠ ${nonCompliant.length} candidates flagged for bias review`); } return { results, complianceRate: results.filter(r => r.euAiActCompliant).length / results.length, meanAIN: results.reduce((s, r) => s + r.ainScore, 0) / results.length }; }
Comparison
ZPL vs Every Other Hiring Approach
Only ZPL provides mathematical proof of fairness. Every other approach leaves you exposed.
Compliance Feature ZPL Layer Manual Hiring Traditional ATS AI Hiring Tools
EU AI Act Mathematical Proof ✅ ✗ ✗ ✗
AIN Score per Candidate ✅ ✗ ✗ ✗
Mathematical Audit Trail ✅ Manual notes only Activity log only ✗
Real-time Bias Monitoring ✅ ✗ ✗ Varies by vendor
No Demographic Data Required ✅ ✗ ✅ ✗
Court-Admissible Evidence ✅ ✗ ✗ ✗
Business Case
The Math on Lawsuit Prevention

One Prevented Lawsuit Pays for ZPL Enterprise for 50 to 333 Years

$300K–$2M
Average cost of a wrongful hiring / employment discrimination lawsuit (legal fees, settlement, reputational damage)
$5,988/yr
ZPL Enterprise plan at $499/month — unlimited compute, 25 API keys, full audit trail export, SLA guarantee
50–333x
Return on investment from a single prevented lawsuit. EU AI Act fines (up to €30M) make the ratio astronomical.
Mathematical proof of fairness changes your legal position from "defendant with no records" to "defendant with a signed, timestamped, court-admissible AIN certificate for every screening decision made." That difference is the entire case.

Hire Fairly. Prove It Mathematically.

Add ZPL to your recruitment pipeline today. Every candidate scored. Every round audited. Every decision defensible.

Get API Key Enterprise Pricing
Questions? Book a call: contact@zeropointlogic.io