Municipal IT helpdesk forbids pasting passwords into prompts and teaches that prompts can leak sensitive data. What prompting basic is emphasized?

Municipal IT helpdesk teaching that prompts can leak sensitive data and forbidding passwords in prompts emphasizes prompt privacy. Anything typed into a generative tool may be logged, retained, or exposed depending on vendor terms and enterprise controls. Credentials, tokens, and similar secrets must never enter municipal AI prompts; helpdesk literacy begins with refusal, not with speed of paste.

Excluding ticket passwords from background in municipal AI prompts describes redaction for context, not the emphasized basic: the teaching is the blanket prompt-privacy ban on placing secrets into prompts, not curating background text to improve answers. Redacting new ticket passwords when preparing municipal AI prompts fails as the emphasized basic: prompt formatting shapes output structure, not input secrecy, and labeling redaction as formatting misnames the prompt-privacy ban on placing secrets into prompts. Instructing helpdesk persona to refuse passwords in municipal AI prompts fails as the emphasized basic: role prompting changes instructions and tone, not actual data handling or retention, and it cannot substitute for never placing secrets into municipal AI prompts.

Exam caveat: prompt privacy covers more than passwords—SSNs, health details, sealed case facts, and API keys belong under the same ban unless an approved, governed tool and redaction process exist. Operational check: publish a one-page "never paste" list, configure DLP where available, train helpdesk on approved reset workflows that keep secrets out of AI chats, and audit sample prompts for credential leakage.