Durham AI Ethics Review and Bias Audit Bylaw Guide

Technology and Data North Carolina 3 Minutes Read ยท published February 09, 2026 Flag of North Carolina

Durham, North Carolina municipal departments that develop or procure AI-powered tools should follow a clear ethics review and bias-audit process to reduce harm, ensure transparency, and meet city legal obligations. This guide summarizes how a Durham-focused review can fit with existing city code, procurement, and oversight structures, explains enforcement pathways, and lists practical steps for departments, vendors, and residents.

Scope and Purpose

This guidance applies to city-developed systems and third-party tools used by City of Durham departments for decision support, automated decision-making, or algorithmic processing of personal data. It covers ethics review triggers, bias-audit expectations, documentation, and reporting. Where the city has enacted specific rules they are cited; where not, the municipal code and procurement rules are the controlling references.City of Durham Code of Ordinances[1]

Start reviews early in procurement and development to avoid costly redesigns.

Key Elements of an AI Ethics Review

  • Document the system purpose, decision points, datasets, and stakeholders.
  • Conduct a documented bias assessment covering input data, model training, evaluation metrics, and potential disparate impacts.
  • Define monitoring and validation plans for model drift, accuracy, and fairness after deployment.
  • Include contract clauses for vendor cooperation on audits, data access, and remedial measures.
  • Set review cadence and triggers for re-audit (e.g., major model update, new dataset ingestion, or complaint).

Penalties & Enforcement

Durham enforces municipal rules through code enforcement, procurement remedies, and where authorized, civil or criminal penalties. For AI-specific fines, escalation, and time limits the city has not published a dedicated AI bylaw on the cited code page; details are therefore not specified on the cited page.[1]

  • Fine amounts: not specified on the cited page.
  • Escalation: first, repeat, and continuing offence ranges are not specified on the cited page.
  • Non-monetary sanctions: the city may use corrective orders, contract remedies, suspension of service, and injunctive relief as available under city code or contract; specific AI sanctions are not specified on the cited page.
  • Enforcer and inspection pathway: enforcement typically involves Code Enforcement, Procurement/Contracts, or the responsible department; complaints can be directed to the city department or official complaint portals listed below.
  • Appeals and review: appeals processes and time limits depend on the controlling ordinance or procurement contract and are not specified for AI matters on the cited code page.
If a vendor refuses data access for audits, procurement remedies and contract termination are primary enforcement tools.

Applications & Forms

No dedicated city form for AI ethics review or bias audits is published on the cited municipal code page; departments should follow existing procurement and contract submission procedures or contact Technology Services for internal review processes.[1]

Action Steps for Departments and Vendors

  • Integrate ethics review into RFPs and statements of work.
  • Require vendors to produce bias-audit reports and documentation of training data provenance.
  • Plan for technical mitigations: reweighting, counterfactual testing, or human-in-the-loop controls.
  • Designate a responsible office or data steward and publish a contact for complaints and inquiries.

FAQ

What triggers an AI ethics review in Durham?
Any procurement or internal development that automates decisions affecting residents, allocates benefits or risks, or processes sensitive personal data should trigger a review.
Who enforces compliance for municipal AI tools?
Enforcement is handled by the department owning the system, Procurement/Contracts, and city code enforcement mechanisms; specific AI enforcement rules are not published on the cited page.[1]
Can residents request audits or appeal automated decisions?
Residents should submit complaints to the responsible department or use the official complaint contact; formal appeal routes depend on the underlying program and are governed by applicable ordinances or contracts.

How-To

  1. Define the system scope and identify decision points requiring review.
  2. Collect and document datasets, model specs, and evaluation metrics.
  3. Run a bias audit using representative test sets and fairness metrics.
  4. Produce a mitigation plan and require vendor commitments in contract language.
  5. Establish monitoring, re-audit triggers, and public reporting where appropriate.

Key Takeaways

  • Embed ethics reviews early in procurement to reduce downstream risk.
  • Require documented bias audits and vendor cooperation in contracts.

Help and Support / Resources

  1. [1] City of Durham Code of Ordinances

Categories

General Governance and Administration Public Safety Public Health and Welfare Land Use and Zoning Housing and Building Standards Transportation Environmental Protection Parks and Public Spaces Business and Consumer Protection Taxation and Finance Labor and Employment Education Civil Rights and Equity Elections and Campaign Finance Events and Special Uses Signs and Advertising Utilities and Infrastructure Technology and Data