A bank is ready to release a credit-explanation agent used in several regions. No one has checked whether the explanations treat applicants consistently across those regions. What should the responsible-AI review require before go-live?
Select an answer to reveal the explanation.
Short Explanation
A letter that sounds fair in one city can land differently in the next. You review the credit explanations for fairness across the regions this agent serves, and you finish that review before go-live. A sentence that tells the agent to be fair is not the review.
Full Explanation
Responsible AI includes fairness, and a credit-explanation agent that serves several regions has to be checked for consistent treatment before it goes live. The review compares how explanations come out across those regions before the agent is released. A longer instruction that tells the agent to be fair is wording, and it produces no evidence about regional outcomes, so it cannot close the gap the review is there to catch. Publishing on one more channel widens who meets the agent and does not compare results across regions, so the fairness question stays unanswered in front of a larger audience. Releasing the same day and writing the comparison afterward means customers see credit explanations while the review is still undone. Exam caveat: the fairness review compares explanations across the regions the agent serves. A longer instruction to be fair, or a note written after release, is not that review. Operational check: run the same credit scenario through the agent for each region in scope and hold go-live until a reviewer has recorded whether the explanations stay consistent.