What If AI Isn’t Making Us Worse at Thinking?
There is a growing anxiety about artificial intelligence and what it may be doing to our ability to think.
Students can ask a chatbot to write an essay.
Workers can summarize documents they have not read.
Search engines increasingly provide answers before we ever click through to the source.
Tasks that once required us to read, compare, synthesize, calculate, or write can now be completed in seconds.
The fear is understandable.
But what if AI is not creating our critical-thinking crisis?
What if .... it is simply making a pre-existing one impossible to ignore?
For decades, many of our institutions have quietly confused access to information with knowledge and producing the correct answer with thinking.
Schools often reward students for remembering what will appear on an exam.
Workplaces reward speed, productivity, and deliverables.
Political discourse rewards certainty.
Social media rewards reaction.
And increasingly, algorithms reward whatever keeps us looking at the screen.
None of these systems necessarily require us to wrestle with an idea, change our minds, interrogate a source, recognize what we do not know, or understand how history and power shape the information in front of us.
Then artificial intelligence arrived and became extremely good at producing many of the things we had been using as evidence that someone had learned.
An essay.
A summary.
A spreadsheet.
A policy memo.
A research question.
A plausible explanation.
A correct answer.
Suddenly, we are forced to confront an uncomfortable question:
If a machine can produce the thing we were measuring, were we measuring thinking in the first place?
Knowing Things Is Not the Same as Knowing How to Think
Information matters.
Facts matter.
Expertise matters.
There is no meaningful critical thinking without something to think about.
But knowledge is more than possessing information, and critical thinking is more than retrieving it.
Thinking requires us to evaluate evidence.
To recognize bias—including our own.
To ask who produced the information and for what purpose.
To distinguish correlation from causation.
To notice who or what is missing.
To understand historical context.
To encounter competing explanations without immediately retreating into the one that makes us most comfortable.
And perhaps most importantly, thinking requires the humility to say:
I might be wrong.
Those capacities were important before generative AI existed.
They are essential now.
AI Is a Mirror
Artificial intelligence is often discussed as though it is an entirely separate intellectual force entering human society.
But AI systems learn from us.
From our books and websites.
Our scholarship and journalism.
Our brilliance and creativity.
Our stereotypes and assumptions.
Our historical records.
Our inequalities.
Our arguments.
Our mistakes.
Our biases.
Our racism.
AI does not arrive untouched by human institutions. It reflects them.
That makes artificial intelligence both extraordinarily powerful and extraordinarily revealing.
Ask an AI system a bad question and it can produce an impressively polished bad answer.
Give it incomplete information and it may confidently fill in the gaps.
Feed it historical data shaped by discrimination and technology can reproduce patterns that look objective because they arrive wrapped in mathematics.
The danger is not simply that people will trust machines too much.
The danger is that we have spent decades building systems in which people were rarely taught how to decide what deserves their trust in the first place.
This Is a Democratic Problem
Critical thinking is often discussed as an educational outcome or workforce skill.
It is also democratic infrastructure.
Democracy assumes people can encounter information and make judgments about it.
That they can evaluate competing claims.
Recognize manipulation.
Understand tradeoffs.
Distinguish disagreement from deception.
Engage with people whose experiences differ from their own.
And participate in decisions that do not come with obvious answers.
That is already difficult in an environment filled with misinformation, fractured media ecosystems, declining trust in institutions, and algorithms designed to maximize attention.
Artificial intelligence raises the stakes.
The answer cannot simply be to ban the technology.
Nor can it be to embrace every new tool because innovation is inevitable.
We need something harder.
We need people capable of using powerful technology without outsourcing their judgment to it.
Education Must Ask a Different Question
Much of the conversation about AI in education has centered on a predictable question:
How do we stop students from using it to cheat?
That question matters.
But it may be considerably less important than another:
What should students be able to do intellectually that a machine cannot do for them?
Perhaps the future of education is not about proving that a student can produce 1,000 words without assistance.
Perhaps it is about whether they can defend an argument.
Interrogate evidence.
Identify what an AI response overlooked.
Compare competing interpretations.
Connect present conditions to history.
Recognize whose perspective is absent.
Ask a better question.
Change their mind when the evidence demands it.
And explain not merely what they believe, but why.
Those are not skills for surviving the AI revolution.
They are skills for participating in a democracy.
The Opportunity Hidden Inside the Crisis
Artificial intelligence may ultimately force us to reconsider what education is actually for.
That could be a gift.
For generations, we have organized enormous portions of education around the transfer, retrieval, and reproduction of information.
Technology is rapidly reducing the scarcity of information.
But information was never the thing democracy needed most.
Democracy needs people capable of making meaning from it.
People capable of determining what is credible.
People capable of seeing beyond their own experience.
People capable of asking who benefits, who decides, and who is missing.
People who can use technology while maintaining the profoundly human responsibilities of judgment, curiosity, empathy, imagination, and moral reasoning.
So perhaps the most important question facing us is not:
Will artificial intelligence learn to think like humans?
It is this:
What if the greatest danger isn't that AI learns to think like us—but that we discover how rarely we were thinking critically in the first place?
That realization should not make us afraid of the future.
It should make us much more ambitious about what we teach people to do with their minds.
Comments