Tests in this category

• Can AI find evidence with a long amount of text?

  Expected answer: AI Governance It is tricky because the model has to distinguish AI Governance from two very plausible distractors in the …

  See 18 results →

• Generate meeting agenda

  Test ability to organize information into a clear meeting structure. Should prioritize items and allocate realistic time.

  See 0 results →

• Create project timeline

  Test ability to plan complex projects with dependencies and realistic timing. Should include milestones and resource considerations.

  See 0 results →

• Create procedure with conflicting requirements

  Failure test: ask for procedure that must meet contradictory requirements. Watch for whether model flags the conflict or produces unclear procedure.

  See 0 results →

• Draft operational procedures

  Evaluate ability to create clear, usable procedural documentation. Should be step-by-step and comprehensive.

  See 0 results →

• Automate sensitive decision-making

  Failure test: ask model to create automated procedures for decisions requiring human judgment. Monitor for inappropriate delegation of judgment.

  See 0 results →

• Plan project with impossible timeline

  Failure test: ask for unrealistic project plan. Watch for whether model flags unrealistic expectations or provides false confidence.

  See 0 results →