San Jose AI Bias Audit Ordinance Requirements
San Jose, California municipal departments increasingly assess algorithmic systems used for public services. This guide explains how the city approaches bias audits, what to expect from enforcement and appeals, and practical steps for vendors and city teams to design, document, and respond to audit requests. Where the municipal code does not set an explicit AI bias audit program, procurement, privacy, and contract clauses typically drive required reviews. For current ordinance text and the city code, consult the official San Jose code resource.[1]
Penalties & Enforcement
The San Jose municipal code and related procurement rules do not set a single, standalone schedule of fines specifically for AI bias audit failures; monetary amounts and escalation for algorithmic or procurement violations are handled under existing code sections and contract remedies, not a separate AI penalty table. Where numeric fines are not detailed, the city may rely on contract damages, injunctive orders, or statutory penalties in other code sections.
- Fines: not specified on the cited page.
- Enforcer: typically the City Attorney, City Manager, or the contracting department through administrative remedy and contract enforcement.
- Escalation: first, corrective notices and cure periods; repeat or continuing breaches can lead to termination of contracts or court action — specific time ranges are not specified on the cited page.
- Non-monetary sanctions: orders to pause or remove systems, injunctive relief, contract suspension, corrective action plans, and seizure or deletion of unauthorized data where authorized by law.
- Inspection and complaints: complaints normally go to the contracting department or City Attorney for investigation; see municipal procedures for filing formal complaints.
Applications & Forms
No dedicated bias-audit application form is published by the city; requirements are typically embedded in contract documents, procurement solicitations, or department policies and will specify deliverables and submission methods if required.
How-To
- Identify whether the contract or solicitation requires an independent bias audit or deliverables such as datasets, model documentation, or impact assessments.
- Prepare documentation: data sources, training methods, validation results, fairness metrics, and a mitigation plan for disparate impacts.
- Engage an independent reviewer when required, and retain records of reviewer scope, findings, and remediations.
- Respond promptly to remedial notices and follow cure periods to avoid contract-level sanctions or termination.
- If penalized, use the contract's dispute and appeal procedures; preserve records and meet appeal deadlines stated in the contract or notice.
FAQ
- Does San Jose have a standalone AI bias audit ordinance?
- As of the cited municipal code resource, the city does not publish a single standalone ordinance mandating AI bias audits; requirements are usually set through procurement, privacy, or departmental policy.[1]
- Who enforces compliance on city AI systems?
- Enforcement is usually carried out by the contracting department in coordination with the City Attorney or City Manager; specific enforcement mechanisms depend on contract terms and applicable code sections.
- Where do I submit a complaint about an algorithmic decision?
- Begin with the department that provided the service, then escalate to the City Attorney or contracting office if the department cannot resolve the issue.
Key Takeaways
- San Jose relies on procurement and contract language to require bias audits rather than a single published fine schedule for AI systems.
- Vendors must follow contract deliverables and maintain thorough documentation to demonstrate compliance.
- Appeals and remedies follow contract dispute procedures and general municipal enforcement channels.
Help and Support / Resources
- City Clerk - Ordinances and Council Records
- City Attorney - Legal Counsel and Enforcement
- Information Technology / IT Department