Animal-control licensing stories include pet-owner addresses in prompt examples for an AI drafting helper. What should the SM/TC coach?
Select an answer to reveal the explanation.
Short Explanation
A stray-pet license card is not prompt fuel. If the model does not need the street number to draft acceptance criteria, leave it off the clipboard. Redact or invent examples; do not teach the tool the owner's house.
Full Explanation
Animal-control licensing stories that include pet-owner addresses in prompt examples need SM/TC coaching to minimize sensitive data in prompts and use redacted or synthetic examples instead. A license card is not prompt fuel: if the model does not need the street number to draft acceptance criteria, leave it off the clipboard. Responsible AI for coach and team drafting includes deliberate PII minimization.
Putting full addresses in every prompt so the model can geocode the work fails because geocoding via an unneeded prompt field is not a justification for leaking residence data. Storing owner addresses in a shared prompt library fails because it multiplies exposure across the ART. Letting the model retain owner addresses because municipal data is "already public" fails because public-record nuances do not authorize casual prompt retention or reuse.
Exam caveat: some municipal records have public elements; that still does not make raw owner addresses default training examples for drafting helpers. Operational check: scrub address fields from prompt templates, replace them with synthetic samples, brief the licensing team on redaction norms, and spot-check upcoming AI-assisted drafts for residual PII before use.