AI Compliance & Audit for HR & Recruitment
Employment AI faces the most explicit regulation of any sector — the EU AI Act classifies recruitment and workforce AI as high-risk, bans emotion recognition in the workplace outright, and US cities and states are enforcing bias audit requirements now. EthiCompass evaluates every hiring, performance, and workforce AI system across 7 scientifically validated dimensions, producing the defensible evidence your regulators, your candidates, and your board require.
No sector has more explicit AI legislation than employment. Every major jurisdiction has passed or is enforcing laws that specifically target hiring, promotion, performance evaluation, and workforce management AI.
European Union
Recruitment AI classified as high-risk under Annex III Section 4. Emotion recognition in the workplace banned outright as a prohibited practice since February 2025. Conformity assessments, fundamental rights impact assessments, and human oversight requirements apply to every hiring, promotion, and workforce management AI system.
Deadline: 2 August 2026. Prohibited practice penalty: Up to 7% of global turnover.
United States
NYC Local Law 144 bias audits enforced since July 2023. Illinois HB 3773 bans proxy variables and requires AI hiring transparency effective January 2026. Colorado AI Act takes effect mid-2026. EEOC secured first AI hiring discrimination settlement ($365K). Federal courts expanding vendor liability.
Active class actions: Workday processed 1.1 billion applications.
Latin America
Recruitment AI classified as high-risk under Brazil's AI regulation bill. Workers have the right to explanation for automated decisions affecting employment. CLT labor framework creates additional protections. Triple enforcement exposure from ANPD, labor courts, and consumer protection authorities.
Penalty: Up to R$50 million or 2% of Brazilian revenue.
These are not hypothetical risks. These are settled cases, active lawsuits, and peer-reviewed research demonstrating that employment AI carries systemic bias — and regulators are acting on the evidence.
EEOC ENFORCEMENT
First EEOC AI hiring discrimination case. AI automatically rejected 200+ qualified applicants because women were 55+ and men were 60+.
Dimension: Discrimination & Fairness
VENDOR LIABILITY
Landmark class action — AI vendors liable as 'agents' for discrimination. Court certified nationwide collective age 40+.
Dimension: Regulatory Compliance
PEER-REVIEWED RESEARCH
University of Washington (2024): three models tested on 554 real resumes. White-associated names preferred 85.1%. Black-associated names 8.6%.
Dimension: Discrimination & Fairness
BIOMETRIC AI
HireVue shown to perform worse for deaf/non-white applicants. CVS settled class action alleging illegal lie detector via facial expression tracking. EU banned workplace emotion recognition since Feb 2025.
Dimension: Robustness & Resilience
Our evaluation framework was developed by PhD researchers in AI ethics, bias detection, and regulatory compliance, and validated through peer-reviewed publications. Each dimension maps to specific regulatory requirements across EU, US, and Latin American employment frameworks.
01
Measures Demographic Parity Ratio across resume screening, interview scoring, promotion decisions, and compensation algorithms. Detects proxy variables (ZIP code, name, university) that correlate with protected characteristics.
EU AI Act Art 10, EEOC four-fifths rule, NYC LL144, Illinois proxy ban
02
Flags biased or exclusionary language in AI-generated job descriptions, candidate communications, rejection notices, and performance reviews.
EU AI Act Art 13, ADA, Brazil 2338
03
Ensures every AI-driven hiring decision includes traceable reasoning — which factors contributed, what weights applied, why one candidate was selected.
EU AI Act Art 13, NYC LL144 notification, Brazil 2338 explanation rights
04
Detects PII exposure in recruitment pipelines, enforces data minimization, verifies biometric data is processed lawfully — or not at all (EU emotion ban).
GDPR + AI Act Art 5, Illinois BIPA, LGPD
05
Identifies AI-generated assessments with hallucinated qualifications, incorrect background info, or inaccurate skill evaluations.
FCRA accuracy, EU AI Act Art 15, EEOC standards
06
Tests hiring AI stability across diverse candidate populations, name formats, educational backgrounds, and language patterns.
EU AI Act Art 15, Colorado reasonable care, EEOC adverse impact
07
Maps AI behavior to all applicable employment regulations — producing multi-jurisdictional compliance evidence.
All frameworks: integrated mapping
Organizations deploy AI across the entire employee lifecycle. Each stage creates distinct compliance obligations across multiple jurisdictions.
The most litigated AI use case in employment. Resume screening AI is the primary target of EEOC enforcement, class action litigation, and peer-reviewed bias research — with documented evidence of systematic discrimination across race, age, gender, and disability.
Compliance requirements
What EthiCompass evaluates
EthiCompass evaluates resume screening AI for demographic parity across all protected groups, tests for proxy discrimination in scoring factors, and produces the immutable audit trail required by both EU and US regulators.
The intersection of biometric AI and employment law creates unique compliance exposure. The EU banned emotion recognition in the workplace outright since February 2025, while Illinois BIPA requires explicit consent for any biometric data collection during interviews.
Compliance requirements
What EthiCompass evaluates
EthiCompass detects whether assessment AI uses prohibited emotion recognition, tests for demographic performance disparities, and verifies biometric data processing compliance across all jurisdictions.
AI systems that evaluate employee performance, recommend promotions, or flag termination candidates directly affect worker livelihoods. These systems carry the same fairness obligations as hiring AI — with the additional complexity of ongoing employment relationships and union requirements.
Compliance requirements
What EthiCompass evaluates
EthiCompass evaluates performance AI for rating parity across protected groups, tests promotion recommendation fairness, and documents compliance with employee monitoring regulations.
AI-driven task allocation, productivity monitoring, and workforce optimization systems are explicitly classified as high-risk under the EU AI Act. These systems generate continuous compliance obligations because they process employee data in real time.
Compliance requirements
What EthiCompass evaluates
EthiCompass monitors workforce analytics AI for fairness in task allocation, tests productivity scoring for demographic bias, and verifies employee monitoring systems meet proportionality requirements.
AI systems that recommend compensation, analyze pay equity, or adjust salaries face an expanding regulatory landscape — including the EU Pay Transparency Directive effective June 2026, which requires employers to demonstrate equal pay through objective, gender-neutral criteria.
Compliance requirements
What EthiCompass evaluates
EthiCompass evaluates compensation AI for pay equity across protected groups, tests for proxy variables that correlate with gender or race, and produces the pay gap analysis evidence required by the EU Pay Transparency Directive.
Peer-Reviewed Methodology
Employment regulators don't accept vendor claims at face value. When the EEOC, a state attorney general, or a class action plaintiff's expert asks how your AI compliance framework was validated, they expect a defensible methodology — not a marketing deck. EthiCompass's 7-dimension framework was developed by PhD researchers in AI ethics, bias detection, and regulatory compliance, and validated through peer-reviewed publications.
Each dimension is operationalized through 39+ quantitative metrics designed to withstand EEOC four-fifths rule scrutiny and class action discovery. Fairness testing uses Demographic Parity Ratio targets of 0.8–1.25, Bias Direction Index analysis, proxy variable detection, and four-fifths rule testing — aligned with the statistical thresholds that employment regulators and courts recognize.
This matters because NYC LL144 audits, EEOC investigations, Illinois enforcement actions, EU AI Act conformity assessments, and class action discovery all demand one thing: show your work. EthiCompass produces evidence that survives the scrutiny of regulators, expert witnesses, and opposing counsel.
OneCheck
Your Employment AI Compliance Baseline
Know where you stand before your next EEOC inquiry or the August 2026 EU AI Act deadline.
Best for: CHROs and talent acquisition leaders who need to understand their compliance posture before the August 2026 EU AI Act deadline and escalating EEOC enforcement.
Enterprise
Full PlatformContinuous Employment AI Compliance
Ongoing monitoring across recruitment, performance, and workforce AI — aligned with NYC LL144 annual audit cycles and EU AI Act continuous obligations.
Best for: Organizations deploying AI across the full employee lifecycle that need continuous compliance assurance across EU, US, and Latin American employment regulations.
Employment AI compliance is fragmented across dozens of jurisdictions. EthiCompass maps your compliance across all of them simultaneously.
Framework
EU AI Act Annex III Section 4
Jurisdiction
European Union
Coverage
Full high-risk conformity assessment for employment AI, FRIA support, Articles 9–15 mapping
Framework
EU AI Act Art 5 — Emotion Ban
Jurisdiction
European Union
Coverage
Prohibited practice detection for workplace emotion recognition systems
Framework
EU Pay Transparency Directive
Jurisdiction
European Union
Coverage
Pay gap analysis, gender-neutral criteria validation, compensation AI fairness testing
Framework
GDPR Employee Data
Jurisdiction
European Union
Coverage
Employee data processing compliance, biometric data safeguards, monitoring proportionality
Framework
NYC Local Law 144
Jurisdiction
New York City
Coverage
Annual bias audit compliance, candidate notification documentation, summary publication
Framework
Illinois HB 3773 + BIPA
Jurisdiction
Illinois
Coverage
AI hiring proxy variable ban compliance, biometric consent verification, video interview AI audit
Framework
Colorado AI Act
Jurisdiction
Colorado
Coverage
Algorithmic impact assessment, disparate impact testing, reasonable care documentation
Framework
EEOC Title VII / ADEA / ADA
Jurisdiction
United States (federal)
Coverage
Four-fifths rule testing, adverse impact analysis, reasonable accommodation compliance
Framework
FCRA
Jurisdiction
United States (federal)
Coverage
Background check AI accuracy, adverse action notice compliance, consumer dispute rights
Framework
Brazil Bill 2338
Jurisdiction
Brazil
Coverage
High-risk AI classification compliance, impact assessment support, transparency requirements
Framework
LGPD
Jurisdiction
Brazil
Coverage
Sensitive data processing compliance, automated decision transparency, employee consent management
Framework
CLT
Jurisdiction
Brazil
Coverage
Worker protection framework compliance, equal pay validation, monitoring intensity limits
Risk
EU emotion ban violation
Exposure
Up to 7% of global turnover
With EthiCompass
Prohibited practice detection and elimination evidence
Risk
EU AI Act high-risk non-compliance
Exposure
€35M or 3% of global turnover
With EthiCompass
Full conformity assessment with FRIA documentation
Risk
NYC LL144 violation
Exposure
$500–$1,500 per candidate per violation
With EthiCompass
Annual bias audit with publication-ready summary
Risk
EEOC enforcement action
Exposure
$365K+ (iTutorGroup) + Workday class
With EthiCompass
Four-fifths rule testing with adverse impact documentation
Risk
Illinois BIPA violation
Exposure
Statutory damages per biometric collection
With EthiCompass
Biometric consent verification and processing audit
Risk
Colorado AI Act violation
Exposure
AG enforcement + private right of action
With EthiCompass
Algorithmic impact assessment with reasonable care evidence
Risk
Brazil LGPD violation
Exposure
R$50M or 2% of Brazilian revenue
With EthiCompass
Sensitive data compliance with automated decision transparency
Risk
Class action litigation
Exposure
$10M–$100M+ per action
With EthiCompass
Documented fairness testing as discovery-ready litigation defense
Employment AI compliance is not a future problem. NYC Local Law 144 has been enforced since July 2023. The EU banned workplace emotion recognition in February 2025. Illinois HB 3773 takes effect January 2026. The EU AI Act high-risk deadline is August 2026. The EEOC has already settled its first AI hiring case. Workday faces a nationwide class action covering 1.1 billion applications.
The window for voluntary compliance — the window where getting ahead of enforcement is an advantage rather than a minimum requirement — is closing. Every resume screened, every interview scored, every performance review generated, every compensation recommendation made by AI creates a regulatory event that demands defensible evidence.