Skip to content

Latest commit

 

History

History
33 lines (27 loc) · 1.56 KB

File metadata and controls

33 lines (27 loc) · 1.56 KB

Roadmap

Phase 0 — Foundation (this release, v1.0)

  • OpenAPI contract implemented end-to-end (openapi/aibf.yaml)
  • Feature model splitting merit vs protected-proxy signals
  • Linear reference model + zero-baseline (SHAP) attributions
  • Bias score, thresholds, plain-language explanations, flagging
  • HR feedback + human-in-the-loop retraining and threshold recalibration
  • Synthetic, deterministic training/demo data (no real PII)
  • Test suite + runnable pipeline demo

Phase 1 — Explainability depth (v1.1)

  • Pluggable non-linear reference models with kernel/tree SHAP.
  • Counterfactual explanations ("would not have been flagged if …").
  • Proxy-variable detection for name/zip-code style leakage.

Phase 2 — Standards mapping (v1.2)

  • Map bias scores to the four-fifths (80%) rule and impact-ratio reporting.
  • Exportable, machine-readable audit reports (JSON + human summary).
  • Alignment notes for NYC Local Law 144 and EU AI Act Article 10 obligations.

Phase 3 — Integrations (v1.3)

  • Adapters to consume real ATS outputs (Greenhouse/Lever/iCIMS-style APIs).
  • Consented, self-reported demographic ingestion (replacing inference).
  • Reviewer UI for the flagged-decision feed.

Phase 4 — Continuous auditing (v2.0)

  • Scheduled re-audits with model-drift and protected-class coverage checks.
  • Dashboards for population-level disparate-impact trends over time.

Non-goals

  • Making legal determinations of discrimination.
  • Shipping or storing real candidate PII in the open-source core.