Fact-check the AI's support and judge the reasoning before you stand behind it.

Challenge

1. Challenge 01

   Audit your steelman: verify every factual claim against a real source, and check the logic for holes (false cause, cherry-pick, smuggled assumption). Separate what is useful from what must be cut. Never blindly trust AI-supplied evidence.

   Show sample answer
2. Challenge 02

   An AI writes: "Social media harms teen mental health because studies show depression rates correlate with screen time." Break apart the reasoning. What claim is being made? What evidence supports it? What assumption is smuggled in? What would need to be true for the conclusion to hold?

   Show sample answer
3. Challenge 03

   The AI generates a paragraph defending a business decision. List every factual claim (names, dates, numbers, specific events). For each, decide: can I verify this myself? Do I have to trust the AI? If you have to trust it, note that it's an assumption, not a proven fact.

   Show sample answer
4. Challenge 04

   An AI proposes a policy by citing three reasons. Rank them by strength of evidence. For the weakest reason, identify what's missing: a statistic? A counterexample? A competing explanation? Rewrite the reason to strengthen it or cut it entirely.

   Show sample answer
5. Challenge 05

   The AI makes a sweeping claim (e.g., "X is the leading cause of Y"). Ask: Leading by what measure? (Raw count, percentage, impact score?) What's the data source? What was left out? Generate two rival explanations the AI didn't consider, then decide if the AI's claim still holds.

   Show sample answer
6. Challenge 06

   Spot the logical gap. The AI writes: "Countries with high renewable energy use have lower carbon emissions. Therefore, renewable energy is the most effective climate solution." What's the gap? What would need to be true for this logic to hold? What counter-evidence might exist?

   Show sample answer
7. Challenge 07

   An AI supports a claim by citing an "expert." Fact-check the expert: What's their field? Are they speaking in their field or outside it? Do they have financial or reputational stakes in the answer? Decide if citing them is fair or if the AI cherry-picked a friendly voice.

   Show sample answer
8. Challenge 08

   The AI uses language like "arguably," "some say," "it's often claimed." Decode these hedges. What's the AI uncertain about? Is the uncertainty honest (the evidence really is thin) or evasive (the AI is hiding a weak spot)? Rewrite the sentence as either a confident claim with evidence or an explicit open question.

   Show sample answer