KG Kinetic Gain Protocol Suite OPEN AI GOVERNANCE SPECS
HR Tech / Employment AI vertical 6-pack

HR Tech / Employment AI
6 sibling specs · 1 canonical shape per artifact

Six sibling specs for AI tools that touch hiring, promotion, performance evaluation, and termination decisions. The Suite vertical 6-pack that lets an employer, AI vendor, or independent bias auditor publish the NYC LL 144 bias-audit + EEOC-investigation-ready evidence that an HR / GC / Chief People Officer chain needs.

See the 6 repos All six verticals →

Regulatory floor

EEOC AI Guidance May 2023 · Title VII §703(k) + UGESP (29 CFR Part 1607) · ADA Title I (42 USC §12112) · Section 504 of Rehabilitation Act · ADEA (29 USC §621) · GINA (42 USC §2000ff) · OFCCP federal-contractor regs · NYC Local Law 144 of 2021 (effective 2023-07-05) · IL Artificial Intelligence Video Interview Act (820 ILCS 42) · MD HB 1202 Facial Recognition · CA AB 331 (withdrawn → DFEH FEHA rulemaking) · CO SB 24-205

Canonical example anchored throughout the 6-pack

Key design innovation vs sibling-vertical equivalents

First audit stream with TWO orthogonal invariants (human-hiring-decision + NYC LL 144 candidate-notice, independent); first Suite tracker supporting sub-state jurisdictions (US-XX-CITY pattern); accommodation-pathway-impairment + two-sd-violation coverage statuses; strictest default-deny posture (6 of 19 vault data categories default NOT a model input).

The six sibling specs

Every Kinetic Gain Protocol Suite vertical 6-pack contains exactly these six artifact shapes. The same six shapes appear in every vertical — only the per-vertical content (data categories, regulatory basis, invariants) differs.

Operator regulatory-lifecycle

state-employment-ai-disclosure-tracker

First Suite tracker supporting sub-state jurisdictions; seed: NYC LL 144, IL 820 ILCS 42, MD HB 1202, CA AB 331, CO SB 24-205

Evidence Bundle (compliance)

eeoc-readiness-evidence-bundle

6 obligation families: Title VII disparate-impact (UGESP) + ADA accommodation + ADEA + GINA + EEOC recordkeeping + OFCCP (conditional)

Evidence Bundle (bias)

employment-candidate-bias-coverage-lab

NYC LL 144 annual bias audit + UGESP four-fifths + UGESP §1607.5(D) 2-SD test + accommodation-pathway-impairment + two-sd-violation

Incident Card

employment-ai-incident-card-profile

13 event types incl. ai-nyc-ll-144-audit-failure + ai-il-820-ilcs-42-violation + ai-md-hb-1202-facial-recognition-violation + ai-ofccp-finding + ai-class-action-plaintiff-filing

Decision Card vault contract

candidate-data-vault-contract-profile

Strictest default-deny: 6/19 categories default NOT a model input (biometric, credit, social, protected-class self-ID, accommodation request, ADA medical); ADA pathway 5-flag block

Why parallel structure matters

A buyer's procurement team operating across mixed regulated verticals — HealthTech AI vendors plus FinTech AI vendors plus HR Tech AI vendors — can apply the same six-shape Suite vocabulary to every vendor in every vertical. The kg-suite-vertical-router tool routes any artifact to the right vertical's verification logic with one CLI command. The kg-suite-vertical-comparator tool surfaces the SAME-vs-DIFFERENT design contributions across all six verticals as a single reference table.