A municipal dog-license tester pastes tonight’s defect notes into a personal chatbot on a home laptop because the shop has not approved any GenAI tool. What term describes using GenAI without formal approval or oversight?
Select an answer to reveal the explanation.
Short Explanation
Think of shadow AI like running an unlicensed food truck behind City Hall — the kitchen might work, but nobody signed off. Pasting defect notes into a personal chatbot without shop approval is exactly that. The label is about missing approval and oversight, not about a prompting trick.
Full Explanation
Shadow AI is the use of generative AI tools without formal organizational approval or oversight. A municipal dog-license tester pasting tonight’s defect notes into a personal chatbot on a home laptop because the shop has not approved any GenAI tool is a clear instance: the channel lacks authorization, logging, and policy control. Prompt injection is a Chapter 3 attack pattern that manipulates model instructions through crafted input—it is not the name for unapproved personal use. Fine-tuning adapts a model with domain-specific training data and is a deliberate specialization technique, not shadow adoption. Retrieval-augmented generation grounds replies with a document store and addresses freshness and relevance, not approval status. Exam caveat: recognizing shadow AI steers testers toward an approved GenAI strategy and deployment path rather than ad-hoc personal helpers for civic defect data. Operational check: if defect notes leave the municipal environment into an unapproved personal chatbot, classify the use as shadow AI and redirect work to sanctioned tools and sanitization practices. Keep civic testware subject to the same review discipline you would apply to human-authored packs.