Part 3: AI in HR & Employment | Module 11

👥 HR AI Applications

AI is transforming human resources across the employment lifecycle, from recruitment to performance management and workforce planning.

Key Application Areas

Resume Screening: Automated filtering and ranking of job applicants
Video Interview Analysis: AI assessment of candidate responses, tone, body language
Skills Assessment: Automated testing and evaluation of candidate competencies
Performance Analytics: AI-driven employee performance evaluation
Workforce Planning: Predictive analytics for staffing and retention
Employee Monitoring: Productivity tracking and behavioral analysis

⚖ Anti-Discrimination Compliance

Employment AI must comply with anti-discrimination laws prohibiting bias based on protected characteristics.

Protected Characteristics by Jurisdiction

Jurisdiction	Key Laws	Protected Characteristics
US Federal	Title VII, ADA, ADEA	Race, color, religion, sex, national origin, disability, age (40+)
EU	Employment Equality Directive	Age, disability, religion, sexual orientation + national laws
UK	Equality Act 2010	Nine protected characteristics including gender reassignment
India	Constitution Art. 14-16	Religion, race, caste, sex, place of birth

⚠ Disparate Impact Risk

Even without intentional discrimination, AI hiring tools can have disparate impact on protected groups. Under US law, employers may be liable if selection rates for protected groups are less than 80% of the rate for the highest group (4/5ths rule).

🏙 NYC Local Law 144 (LL144)

New York City's Local Law 144 is a landmark regulation requiring bias audits of automated employment decision tools (AEDTs).

Key Requirements

Bias Audit: Independent audit within one year before use
Summary Publication: Publish audit results on employer website
Notice to Candidates: Notify candidates 10 business days before use
Alternative Process: Offer alternative selection process upon request
Data Disclosure: Inform candidates about data retention

Bias Audit Metrics

Metric	Description
Impact Ratio	Selection rate for category / selection rate for most selected category
Categories	Sex (male, female), Race/ethnicity (as per EEOC categories)
Intersectionality	Impact ratios for sex categories within each race/ethnicity category

🇪🇺 EU AI Act on Employment AI

The EU AI Act classifies AI systems for employment purposes as high-risk, requiring compliance with stringent requirements.

High-Risk Employment AI (Annex III)

AI for recruitment and selection of candidates
AI for making decisions affecting terms of work relationships
AI for task allocation based on individual behavior or traits
AI for monitoring and evaluating performance and behavior

Compliance Requirements

✓Risk management system implementation
✓Data governance and quality management
✓Technical documentation
✓Record-keeping and logging
✓Transparency and information to users
✓Human oversight measures
✓Accuracy, robustness, cybersecurity

📊 Best Practices for HR AI Governance

✓Conduct bias audits before deployment and annually thereafter
✓Document the business necessity for AI-based criteria
✓Ensure meaningful human review of AI recommendations
✓Provide clear notice to candidates about AI use
✓Maintain records for potential discrimination claims
✓Review vendor contracts for compliance obligations

💡 Vendor Due Diligence

When using third-party HR AI tools, employers remain legally responsible for discrimination. Conduct thorough vendor due diligence including review of bias testing, audit reports, and contractual allocation of compliance responsibilities.

📚 Key Takeaways

1Employment AI is subject to anti-discrimination laws across all jurisdictions
2NYC LL144 requires bias audits and transparency for automated hiring tools
3EU AI Act classifies employment AI as high-risk with extensive requirements
4Disparate impact liability exists even without intentional discrimination
5Employers remain responsible for vendor AI tool compliance