A FAQ chatbot is expanding to give preliminary eligibility guidance for housing assistance. What governance action should accompany that scale step?
Select an answer to reveal the explanation.
Short Explanation
Moving from “where’s the form?” to “might you qualify?” is a bigger leap than a UI tweak. Re-tier the risk before you scale into eligibility—stakes and safeguards change together.
Full Explanation
Expanding a FAQ chatbot into preliminary eligibility guidance for housing assistance raises harm and compliance stakes: wrong guidance can affect benefits access, create reliance interests, and trigger fairness, privacy, and appeal concerns. Governance should reassess and re-tier risk, strengthen controls, and often increase human review before that scale step rather than reusing the prior informational FAQ tier. Keeping the original FAQ risk tier unchanged because the same vendor model is used confuses model brand with use-case impact; risk follows decision consequences, not the logo. Skipping governance review to accelerate the housing department’s launch date trades schedule for unmanaged high-stakes errors. Treating eligibility guidance as lower risk than FAQ because fewer pages are involved mistakes content volume for consequence severity—fewer pages can still determine whether a resident seeks aid. Exam caveat: CAIPM-style items emphasize re-tiering when AI moves from informational to eligibility-adjacent guidance; same model does not equal same risk. Operational check: run a fresh risk assessment, update DPIA/bias and appeals considerations, require human confirmation or clear non-determination disclaimers as policy dictates, expand evaluation sets for eligibility edge cases, and obtain governance approval before enabling the higher-stakes scope.