Artificial intelligence is already embedded in municipal communication—from live captions and translations to drafting notices and triaging constituent emails. For city and county clerks, the ethical obligations are concrete: be accurate, be fair, be transparent, safeguard records, and protect residents’ privacy. This white paper offers a practical ethics framework that translates these obligations into console-visible controls, procurement language, and publication practices that residents can see and verify.

Three themes guide the next five years. First, ‘explainable enough’ operations: staff and the public must be able to understand why an AI system produced a particular output and how to correct it. Second, ‘proportional’ use: higher-salience decisions (e.g., notices affecting rights) require tighter human review, auditability, and public documentation than routine summarization. Third, ‘portability and accountability’: all artifacts—captions, translations, prompts, logs—must be exportable, retained per schedule, and open to audit without vendor permission.

We provide a tiered risk model, a standardized disclosure policy, a bias and harm review workflow, and concrete procurement clauses. Tables and checklists map ethical principles to what operators and residents actually feel: clear audio and text, correct language usage, visible corrections when needed, and a complete, trustworthy archive. Throughout, we favor controls that are observable by staff and legible to the public over abstract assurances or vendor claims.

The paper closes with a two-year roadmap for small teams: establish disclosure and consent patterns, stabilize accessibility as an SLO, implement exportable logging, and build a cadence of public-facing correction notes. In doing so, clerks can meet statutory obligations, build resident trust, and benefit from AI without ceding control over the public record.