After hours, a community book-bike tester pastes a live checkout outage into an unapproved chat pane “just to draft a defect report faster.” How should this situation be classified?
Select an answer to reveal the explanation.
Short Explanation
Speed does not turn an unapproved helper into a shop tool. Pasting a live checkout outage into a personal chat pane to draft a defect faster is shadow AI. It is not a Chapter 2 prompting technique and it is not CT-AI testing of the model.
Full Explanation
A typical test-shop instance of shadow AI is using an unapproved GenAI pane on production defect material without formal approval or oversight. After hours, a community book-bike tester pasting a live checkout outage into an unapproved chat pane “just to draft a defect report faster” matches that pattern: live production content enters an unsanctioned channel. Chain-of-thought prompting describes how to structure reasoning in a prompt, not whether use is authorized. Few-shot prompting with production incident examples is a prompting technique and still does not authorize an unapproved pane. Metamorphic testing of the chatbot as the system under test belongs to testing AI systems, not to classifying unapproved GenAI-for-testing use. Exam caveat: prompting technique names and CT-AI SUT methods must not be substituted for the shadow-AI definition. Operational check: if live civic checkout outage details are about to enter a chat pane, confirm the tool is approved and logged; otherwise stop and use sanctioned defect-reporting channels. Keep civic testware subject to the same review discipline you would apply to human-authored packs. On the written exam, prefer syllabus terminology over vendor marketing language or lab choreography.