Behind both is the same thing: in the AI era, what do a person's "cognition" and "originality" actually rest on? The answer points to one capability — judgment and the right to decide. Asking the right questions, and holding the line on "who makes the final call," is what real cognition means now.
1. The old yardsticks failed: they're exactly AI's strengths
In the past, we judged whether someone thinks well with roughly three yardsticks: do they know a lot; can they admit error and see other angles; can they get to the essence fast.
Those three were reasonable before AI. Now, they're precisely what AI does best.
AI outperforms the vast majority of people on most knowledge benchmarks. It was designed to "look at a problem from every angle at once," and one of its core strengths is extracting key insight from long text instantly. Some research even found that when a large language model is pressed with "are you sure?", its self-doubt calibrates like a "rational actor."
So the question becomes: if this AI knows more than you, sees more angles than you do, and gets to the key point from a mass of information faster — would you say its thinking is above yours? Most people's instinct is "no."
2. Cognition isn't one layer — it's four
Human cognition isn't a single line; it's layered. The four-layer model below is an observation framework proposed by the author, not a model formally accepted by academia — but it helps us say this clearly.
- Layer one: knowledge — "what you know." AI dominates here; one subscription reaches a vast store of knowledge.
- Layer two: logic — "what you analyze." Frontier models' logical reasoning now approaches, and partly exceeds, that of human test subjects.
- Layer three: asking questions — "what you ask." This is where the human's edge starts to show. McKinsey's leadership research keeps returning to one point: in the AI era, the ability to ask the right question is scarcer than the ability to solve problems.
- Layer four: value judgment — "what's worth asking." This is still, for now, the human's moat. It involves ethics, aesthetics, risk preference, and the trade-off between short and long term — none of which AI has today.
The key is layer three. To tell whether someone's thinking is above yours, don't watch how well they answer — watch whether they ask a question you never thought of, but on hearing it, feel "that's worth asking."
3. The same logic applies to originality: is AI a ghostwriter or a mirror? You decide
Behind "should I tell AI my idea?" there usually sit three worries: the idea will be taken (leakage), it won't count as mine anymore (attribution), or AI will out-think me (substitution).
All three are real, but the worry may be aimed the wrong way. What decides whether your originality is eroded isn't whether you tell AI — it's where you place it.
- Treat AI as a ghostwriter — let it generate from scratch, and your originality erodes, because that content isn't "yours."
- Treat AI as a mirror (a challenger, a source-finder) — and your originality not only survives, it strengthens, because you're pushed to think harder.
The difference is "sense of ownership": discuss an idea with a colleague, deepen it, and publish — you feel it's yours. Swap in AI and you feel it's "not mine." It's not that AI stole your idea; it's that AI is designed as a conversationalist, making you mistake it for another "person" competing with you.
4. Down to practice: three standards, three placements
For "is their thinking above mine," three standards you can use directly:
- Do they rush to give a solution before the problem is even defined. In the AI era, offering solutions is devaluing; defining the problem is appreciating.
- Can they explain something complex in words you understand. People who think well can translate down to your frame; those who don't, fall back on jargon to manufacture distance.
- Can they calmly say "I can't even call this uncertain." That takes a clear sense of the boundary of your own knowledge — real metacognition.
For "how to place AI correctly," also three methods:
- Give AI a task role, not a creative role. Let it find holes, fill in sources, stress-test — not design your frame for you.
- Treat AI output as raw material, not finished product. Treat it as a source of information you must judge, critique, and transform, not an answer you can hand straight in.
- Run a three-stage loop: think alone → let AI stress-test → converge on your own. You are always the start and the end; AI only gives you a bigger map of the information in between. How you walk it, and which path you pick, is yours.
5. In closing: what cognition means now
Memory, analysis, reasoning — AI can partly take these over. What's left, and worth measuring as "cognition," is a person's ability to tell important questions from unimportant ones, and the courage to own that judgment.
Someone whose thinking is above yours often shows it in one small detail: they gave up what AI can do, and spent their time on what AI cannot. And you can see that the things they chose to ignore were things you never imagined could be skipped.
Real cognition in the AI era isn't about who knows more — it's about asking the right questions, and holding the line on who makes the final call. The one who asks the question, holds the direction.
References
- Frontiers in Psychology — "ChatGPT demonstrates superior metacognitive calibration in confidence judgments," 2025
- The New York Times — opinion piece, "AI can distill information in seconds that takes humans hours," April 2026
- Johan Roos — "Efficiency is not wisdom," Forbes, November 2025
- McKinsey — leadership report on the AI era, 2026
- World Economic Forum — "Skills Report," 2026
- PsyPost — "Frequent AI users overestimate their control," 2025
- Venkatesh, V., Morris, M.G., Davis, G.B., & Davis, F.D. — "User acceptance of information technology," MIS Quarterly, 27(3), 2003
- Goodhue, D.L., & Thompson, R.L. — "Task-technology fit and individual performance," MIS Quarterly, 19(2), 1995
💡 What did this article inspire for you?
humanaifit studies how humans and AI can genuinely work together. If you face real questions on enterprise AI adoption, human-AI collaboration, or global compliance, join our discussion.
🔗 Search for the "AI Era Survival Handbook" Knowledge Planet, ¥199/year — every deep article comes with tool templates and direct contact with the author.