A records retention office is evaluating vendors for a generative AI records assistant. How should procurement support AI governance?
Select an answer to reveal the explanation.
Short Explanation
Vendor questionnaires are where city AI principles grow teeth—privacy, security, retention, no surprise training on your records. A shiny demo is not a substitute for answers in writing.
Full Explanation
Procurement supports AI governance for a records retention office evaluating generative AI vendors by aligning vendor questionnaires with city AI principles on privacy, security, and retention. Written questions force vendors to state data use, training on municipal records, subprocessors, encryption, retention and deletion, breach duties, and residency. Mapping those answers to adopted municipal AI principles turns ethics into contract-selectable criteria rather than slideware presented only during demos.
Selecting the cheapest vendor without asking about data use or retention optimizes price while ignoring the core governance risks of a records assistant. Skipping security questions if the sales demo looks polished substitutes theater for assurance. Allowing vendors to train freely on all municipal records by default may permanently encode sensitive civic content into vendor models outside city control.
Exam caveat: questionnaires support selection and contracting; they do not replace ongoing vendor risk management after award. Operational check: include AI-principle-mapped questions in the RFP or RFI, score responses explicitly, require contractual bans or limits on training on city records, retain answers with the procurement file, and re-ask material questions at renewal.