A city assurance lead briefs council on AI oversight. Which definition best captures AI assurance?
Select an answer to reveal the explanation.
Short Explanation
Assurance isn’t a shiny sticker you slap on and forget. It’s the ongoing proof pile—valid, fair, secure, compliant—so council can see the system still deserves trust.
Full Explanation
When a city assurance lead briefs council, AI assurance is best defined as the ongoing collection of evidence that systems remain valid, fair, secure, and compliant. That definition ties oversight to measurable controls across the lifecycle—not to a purchase event or a launch promise—and gives council a durable basis for trust, funding, and remediation decisions about municipal AI serving residents.
A one-time marketing badge purchased from any vendor website fails because badges are not continuous evidence and can be unrelated to the city’s actual controls or operating environment. A promise that models never need monitoring after launch day contradicts drift, abuse, data-shift, and policy-change realities in production civic services. An annual slide deck with no testing, evidence, or remediation is presentation without assurance substance and cannot support council fiduciary oversight.
Exam caveat: assurance is an evidence program over time; certification theater or launch-day optimism do not substitute for ongoing validity, fairness, security, and compliance proof. Operational check: for each in-scope AI system, confirm council-facing packs include current test results, monitoring summaries, open findings with owners, and remediation status—not logos or unread decks alone.