πŸ›‘οΈ Sector Evidence Hubs:
Sector Evidence Hub: hr.aiacts.app

Workforce & HR AI Employment Governance Hub

Why is AI in recruitment and employee evaluation classified as high-risk?

AI systems used in recruitment, job screening, task allocation, and performance monitoring are classified as High-Risk under EU AI Act Annex III point 4, mandating bias prevention, explicit human override controls, and NYC Local Law 144 compliance.

8-Dimension Evidence-Requirement Matrix (Annex III)
01 Model Governance

Candidate Screening Model Validation

Job-related competency validation logs (Annex III 4(a))

02 Fairness

Four-Fifths Rule & Disparate Impact Audit

Statistical bias audit on protected demographic categories (NYC LL 144)

03 Human-in-the-loop

HR Recruiter Human-in-the-Loop Override

Certified human review required for all candidate rejections

04 Transparency

Candidate Transparency & Automated Decision Notice

Pre-assessment notification & disclosure of automated evaluation criteria

05 Data Governance

Resume Corpus De-Identification & PII Masking

Removal of demographic proxies (names, addresses, graduation years)

06 Continuous Monitoring

Hiring Funnel Selection Rate Drift Tracking

Quarterly candidate throughput & demographic parity monitoring

07 Resilience

Anti-Gaming & Resume Injection Defense

Detection of invisible prompt injections & adversarial resume formatting

08 Evidence Packaging

Cryptographic Independent Bias Audit Certificate

Annual public bias audit summary & verifiable SEAL passport

Sector AI Models & Benchmarks (Hugging Face / Open Source)
+ Submit Sector Model β†’

Foundation models, fine-tunes, and specialized architectures classified for this vertical regime.

Google DeepMind Tier 1 GPAI
Gemini 2.5 Pro

Gemini 2.5 Pro surpasses the 10^25 FLOPs threshold, classifying it as a GPAI model with systemic risk under Article 51. Obligations include mandatory model evaluations, adversarial testing, systemic risk assessments, cybersecurity safeguards, and technical documentation for downstream deployers.

View Assessment β†’ Claim Model β†’
Google DeepMind Standard
Gemini 2.5 Flash

Gemini 2.5 Flash operates below the 10^25 FLOPs cumulative computation threshold, classifying as standard General Purpose AI (GPAI). It must maintain technical documentation, comply with EU copyright laws, and supply transparency summaries without systemic risk mitigation burdens.

View Assessment β†’ Claim Model β†’
Google DeepMind Tier 1 GPAI
Gemini 3.7 Flash

Gemini 3.7 Flash includes native support for deterministic human-in-the-loop gating when orchestrated via MCP tool calling, allowing downstream providers to fulfill Article 14 oversight requirements in automated decision workflows.

View Assessment β†’ Claim Model β†’
Audited Sector Training & RAG Datasets (Art. 10)
+ Register Dataset β†’
Art. 10 Provenance Audited
FineWeb (Hugging Face)

undefined

Inspect Dossier β†’ Claim Dataset β†’
Art. 10 Provenance Audited
Common Crawl Corpus

undefined

Inspect Dossier β†’ Claim Dataset β†’
Art. 10 Provenance Audited
The Pile (EleutherAI)

undefined

Inspect Dossier β†’ Claim Dataset β†’
Interactive Sector Spaces & Sandboxes (Art. 57)
Explore All Spaces β†’
⚑ VOS Edge / Workers ⭐ 142
EU AI Act Statutory Explorer

Interactive Article-by-Article statutory explorer, obligations matrix, compliance timeline, and penalty calculator.

Launch Space β†—
⚑ Cloudflare Workers / D1 ⭐ 98
InsureAI Underwriting Sandbox

Live insurance risk sandbox testing protected class proxy discrimination, actuarial fairness, and Article 86 denial explainability.

Launch Space β†—
Designated Statutory Authorities & Regulators
European Labour Authority (ELA) Active
Workforce Rights & Automated Decision Oversight
National Data Protection Authorities Active
Worker Surveillance & GDPR Article 88 Supervision
EEOC (US) Active
Uniform Guidelines on Employee Selection Procedures