A municipal IT strategy committee wants to reduce exposure of housing and benefits case notes to consumer AI services. Which strategy choice best supports that goal?

Municipal strategy committees reduce exposure of housing and benefits case notes by choosing where those notes may be processed, not by hoping consumer tools will treat civic confidentiality as a default. Preferring private-tenant or city-controlled AI environments over public consumer chatbots keeps casework content inside contractual, residency, retention, and access boundaries the municipality can specify and audit. That choice supports confidentiality for residents while still allowing summarization and drafting productivity under governed conditions. Encouraging free public chatbots to avoid licensing cost inverts the risk calculus: consumer services typically train, log, or retain prompts under terms unsuitable for benefits and housing records, expanding rather than reducing exposure. Requiring every case note on the open-data portal before AI use publishes protected civic content and defeats confidentiality goals outright. Disabling all prompt logging removes forensic and oversight capability without stopping data from leaving municipal control; it hides misuse instead of containing it. Exam caveat: private tenancy is not automatic compliance—contracts, DLP, identity integration, and retention still must match the sensitivity of case notes. Operational check: verify caseworker AI routes to approved private endpoints, block consumer chatbot domains for casework roles, and sample prompts to confirm housing and benefits identifiers are not leaving city-controlled environments.