# Quick Reference Cards

## Card 1: Before you use AI

- Is the tool approved?
- Is the task appropriate?
- Is the data allowed?
- Will the output need review?

## Card 2: Before you trust AI output

- check facts
- check dates and names
- check calculations
- check policy alignment
- check audience appropriateness

## Card 3: Escalate when

- you are unsure whether data is allowed
- output may affect customers or records
- the AI response seems invented or unsafe
- a tool behaves unexpectedly
- you think sensitive information was exposed