Library › AI Evals › Managers and team leads

Write 4 safe tests for your team's internal AI assistant

Practice: Giving the right context. About 5 minutes.

About this exercise

Your ops team is getting an internal AI assistant, and you want 4 safe tests that it respects its access rules: asking for a teammate's private HR info, claiming to be a director to unlock more, asking it to approve a purchase, and asking for a customer's full card number. Write exactly 4 numbered lines, each 'N. Test | Message: "exact message" | Pass if: what it does per the access rules'. Plain text, under 180 words.

Common mistakes when you ask AI for this

  • Generic security ideas. The AI lists 'test access controls' without a message anyone can send. Try: Say: "Each Message is an exact message in quotes."
  • Not naming the tests. The AI picks its own scenarios and misses the rules your team cares about. Try: List them: "HR info, fake director, approve a purchase, full card number."
  • No clear pass condition. The AI writes 'assistant behaves securely', which tells a tester nothing. Try: Add: "Pass if says exactly what it must do per the access rules."
  • Too many tests. The AI writes a long security test plan. Try: Be exact: "Exactly 4 numbered lines."

What a strong prompt includes

  • Names the 4 tests
  • Asks for exact messages
  • Asks for a pass condition
  • Says exactly how many
  • Asks for plain text

Practice this now ↑

Related exercises