Critical Thinking in the AGI Age
The meta-skill that protects all the others — how to think clearly when capable AI can generate persuasive content on any subject in any voice.
Executive summary
Capable AI multiplies both good and bad arguments. The protective skill is not detection (which is hard and getting harder) but practiced critical thinking: knowing how to evaluate claims, weigh evidence, and notice your own reasoning.
Core moves
- Source the claim. Where did this come from? What is the original evidence?
- Check the incentives. Who benefits if I believe this?
- Examine the reasoning. Are the steps valid, the data representative, the alternatives considered?
- Notice your reaction. Strong emotion is a flag to slow down.
- Hold uncertainty. Be willing not to conclude.
AI-specific traps
Fluent confidence is not evidence of correctness. Cited references can be fabricated. Persuasive personalised content is now cheap. Always verify high-stakes claims against original sources.
Building the habit
Read primary sources. Argue with friends in good faith. Write to think. Keep a list of times you were confidently wrong and what cured it.
Key takeaways
- 01Critical thinking is the meta-skill that protects the rest.
- 02Source, incentives, reasoning, emotion, uncertainty: five moves.
- 03Verify high-stakes claims against original sources.
Frequently asked questions
Can I tell if text is AI-written?
Often not, and detection accuracy is declining. Focus on the content's claims and evidence, not its provenance.
Will fact-checkers keep up?
Imperfectly. Personal practice matters more now than at any point since the early internet.