[ Use AI Securely ]

[ Chapter 6 · Conclusion and next steps ]

Your personal AI checklist

3 min read

Print this lesson, screenshot it, or copy it into your notes. It is the whole course in ninety seconds.

Before you type

  • Am I in a tool my organisation approved, on my work account?
  • Paste test: would I email this content to an outside company without a contract? If not: placeholders, abstraction, or a synthetic example.
  • Does this involve personal data about customers, colleagues, or candidates? If yes, only in a tool approved for that, if at all.

Before you use the answer

  • Which claims are load-bearing? Verify those against an independent source.
  • Did the model supply facts from its own memory (numbers, names, dates, citations, legal or medical statements)? Those are the hallucination zone: check every one that matters.
  • What tier is this? Internal skim, travelling document, or expert-review territory?
  • Is anything here a decision about a person? Then a human decides, and I could defend that decision without mentioning the tool.

Before you send

  • Would the recipient be misled about who or what produced this? If authorship matters, say how AI was involved.
  • For consequential work: did I keep a one-line trace (tool, input shape, what I verified)?
  • Gut check: if this turns out wrong, is my answer "here was my process" or "the AI did it"? Only the first is acceptable.