San Diego AI Ethics & Bias Audit Ordinance Guide
San Diego, California municipal programs increasingly use automated decision systems and AI. This guide explains how city bylaws and official procedures apply to AI ethics and bias audits for municipal use in San Diego, identifies responsible departments, and outlines steps for compliance, reporting, and appeals. Where the City has not published a single dedicated AI ordinance, this article points to the closest official sources for municipal code, procurement, privacy, and IT policy and explains what is specified and what is not on those pages.
Scope & Definitions
This guidance covers municipal use of automated decision systems, algorithmic tools, and AI models used by city departments for public services, permitting, enforcement, or internal decision-making. Key terms include "automated decision system", "bias audit", "training data", and "affected class" as used in procurement and privacy contexts. For the controlling municipal code and general city rules, consult the City of San Diego Municipal Code.City of San Diego Municipal Code[1]
Standards & Audit Requirements
As of the cited official pages, San Diego does not publish a single consolidated AI ordinance; municipal departments apply procurement, privacy, nondiscrimination, and records rules to AI projects. Departments typically require documented risk assessments, data lineage, and vendor attestations when procuring systems that affect individuals. For department-level policy and data governance, consult the City IT and privacy pages.City of San Diego IT[2]
- Require documented bias audits before deployment.
- Maintain procurement records and vendor documentation.
- Preserve training data inventories and data minimization logs.
- Apply nondiscrimination and civil rights reviews to outcomes.
Penalties & Enforcement
San Diego enforces municipal code, procurement, and privacy rules through department review, contract remedies, and civil enforcement where applicable; specific fine amounts for AI ethics or bias audit failures are not stated on the cited municipal pages and must be determined from contract terms or applicable code sections.[1]
- Monetary fines: not specified on the cited page; amounts depend on the applicable code section or contract remedies.[1]
- Escalation: first, repeat, and continuing offence treatment is not specified on the cited page and may follow civil enforcement or contract termination provisions.[1]
- Non-monetary sanctions: issuance of corrective orders, suspension of system use, contract termination, injunctions, or referral to the City Attorney for civil action.
- Enforcer and inspection: responsible offices include the procuring department, City Attorney, and City Auditor; complaints route through departmental contacts or the City Clerk for formal filings.[1]
- Appeals/review: appeal paths are governed by the underlying code or contract terms; specific time limits are not specified on the cited page and must be confirmed in the controlling instrument or department notice.[1]
Applications & Forms
There is no single city form titled for "AI bias audit" published on the cited pages; departments generally accept audit reports, procurement attachments, and vendor declarations as part of solicitation or contract packages. Where a department publishes a required form, it appears on that department's procurement or IT policy page.[2]
- Submit bias audit reports and supporting documents as attachments to procurement submissions or to the responsible department's records system.
- Contact the procuring department for any specific submission method or required form.
Action Steps for Departments
- Inventory AI systems and classify risk levels.
- Require third-party or internal bias audits before public deployment.
- Include audit deliverables in procurement contracts and SLAs.
- Establish complaint channels with City Clerk and preserve records for review.
FAQ
- Does San Diego have a dedicated AI ordinance requiring bias audits?
- Not currently; the City has not published a single, consolidated AI ordinance on the cited municipal pages and governance is applied through procurement, privacy, and nondiscrimination rules.[1]
- Who enforces compliance for AI systems used by the city?
- Enforcement is typically by the procuring department, City Attorney, or City Auditor; formal complaints may be filed with the City Clerk for review.[1]
- Where do I submit an AI bias audit or complaint?
- Submit audit reports as procurement attachments to the responsible department; for complaints contact the City Clerk or the relevant department's published contact page.[2]
How-To
How to perform a municipal AI bias audit for a San Diego program:
- Define scope and affected populations and document purpose.
- Collect data inventories, training sets, and model documentation.
- Run technical bias tests and validate results with disaggregated metrics.
- Produce a written audit report with remediation steps and submit to procurement/IT.
- Monitor post-deployment outcomes and schedule periodic re-audits.
Key Takeaways
- San Diego currently applies existing procurement, privacy, and nondiscrimination rules to AI systems.
- Specific fines or time limits for AI audit failures are not specified on the cited municipal pages.
- Contact the procuring department, City Attorney, or City Clerk for enforcement, forms, or appeals.
Help and Support / Resources
- City Attorney, City of San Diego
- City Clerk, City of San Diego
- Development Services, City of San Diego