AI and young people: cognitive foreclosure
by Miguel Lucas
The problem with young people using AI to do their schoolwork isn’t that they’re cheating. The real problem is that they may never learn to think. Is there anything we can do about it?
A few days ago, representing the José Antonio Llorente Foundation, I stood in front of hundreds of teenagers at a Talent Session organized by the Princess of Girona Foundation, talking about AI and employability. I dropped a statistic: 83% of students who use ChatGPT to write a paper cannot provide a single correct citation from those texts a week later 1. Not one. I asked them: is that learning, or is that theater?
Cognitive science has a name for this: “cognitive foreclosure” 2. A young person who delegates a task they haven’t yet learned to do independently doesn’t lose a skill. They never build it. It’s not a muscle that atrophies. It’s a muscle that never forms at all. And that weakness can have permanent consequences for reasoning capacity and the development of an intellectual identity.
I didn’t tell those kids to stop using AI (that would be like banning calculators). I told them the same thing I tell my own children: the key isn’t to stop using it — it’s to learn to get value from it without letting it hollow you out. The difference fits in a metaphor: AI cannot be your secretary. It has to be your sparring partner. The training partner who hits back, who forces you to think faster, to defend yourself better. If you only punch the bag, you never improve. You need someone who challenges you. Here are two of the principles I offered them:
-
Don’t ask it for the answer. Ask it to give you better questions. Instead of “write me an essay on climate change,” try: “I want to argue that [X]. Give me the three strongest counterarguments so I can get ahead of them.” That’s using AI to think MORE, not less. The effort of rebutting those counterarguments is yours. And that effort is exactly what builds the muscle.
-
Find the AI’s mistakes. Sometimes it invents data with complete confidence. It cites studies that don’t exist. It builds arguments that sound airtight but, when you pull the thread, rest on nothing. Your job is to catch it. Read every response with suspicion, verify what doesn’t add up, dismantle what seems too perfect. Every error you spot strengthens the one thing no algorithm can give you: your own judgment.
Because AI is an amplifier. Perhaps the most powerful one humanity has ever invented. Plug laziness into it, and it produces world-class idleness. Plug emptiness into it, and it returns that emptiness with excellent prose. But plug in curiosity, and it produces minds more brilliant than any previous generation could have been. Plug in authenticity and judgment, and it will amplify the one thing AI can never produce on its own. Our children already have the amplifier switched on. What we teach them to plug into it will define their future.
Related theses
- Thesis 13 Human judgment will remain a differential value. And AI, the greatest threat to its eventual loss.
- Thesis 19 The danger is not that the machine fails. It is that when it fails, we will have forgotten how to think without it.
- Thesis 22 AI does not make you think better. It makes you feel that you think better.