Is AI making us think less?
It can, if we let it replace our thinking rather than support it. Early research suggests that relying heavily on AI is linked to weaker critical thinking, through something called cognitive offloading. But a lot depends on how we use it.
I'm a fully grown adult, I study this for a living, and I'm already noticing it in myself. I still work on my relationship with my phone and my fragmented attention. And now I catch myself using AI, instead of my own brain, to answer the simplest of questions.
If I'm noticing that, what does it mean for young people whose brains are still developing?
What is cognitive offloading?
Cognitive offloading is handing mental work over to something outside ourselves. We've always done it. We write shopping lists, use calculators, and let our phones remember numbers and directions.
What's different now is that we can offload the thinking itself: working something out, forming an argument, finding the words. And AI is so quick and so capable that it's very easy to offload far more than we mean to.
What the research says so far
"Cognitive debt"
Researchers at MIT Media Lab measured brain activity while people wrote essays, some using ChatGPT and some on their own. Those using ChatGPT showed the lowest brain engagement, and most struggled to recall what they had just written. When they were later asked to write without it, their brains stayed less engaged. The researchers called this "cognitive debt": AI saves mental effort in the short term, but builds up costs over time.
Critical thinking
A 2025 study by Michael Gerlich found that people who used AI tools more often tended to score lower on critical thinking. Cognitive offloading was the key link, and younger people, aged 17 to 25, were the most affected.
Reading on screens
Separately, a large review of reading research found that people tend to understand what they read better on paper than on screens. Worth knowing as more learning moves onto devices.
These are early findings. The evidence is conflicting, studies are often small, and we still lack long-term insights. The MIT study hasn't yet been peer reviewed, and the Gerlich study shows a link rather than proving cause. Neither explores the interesting middle ground: what happens when AI is used to support and deepen thinking, rather than simply replace it. But together, they raise questions that schools and parents need to be asking.
Further reading: Beware of metacognitive laziness, Fan and colleagues, 2024.
What young people are saying
Young people are more aware of this than we often assume. In research with young people across the UK by the Ada Lovelace Institute and the Nuffield Foundation, 'I love it, but I hate it', most worried that AI would make people "intellectually lazy" and stop them thinking for themselves. They were clear that AI should complement human skills, not replace them.
One young person in Islington put it like this:
"It kind of makes you… intellectually lazy. But… we can work well with AI together."
What I see in classrooms
When I delivered The Attention Experiment at Sherborne School, I taught a session to Year 10 boys on what happens in the brain when you focus on a task without interruption, compared with what happens when you break off to check a message or switch apps.
It's simple, foundational knowledge. And it's largely missing. I don't see young people being taught enough about cognitive offloading, how attention works, or what it means to outsource their thinking. Nor are they being helped to understand what's happening when they talk to a chatbot.
This matters all the more because young people are forming relationships with AI, not just using it for homework. More on AI companions
Fast brain, slow brain
One of the things tech does is activate our fast brain. A lot.
But we use our slow brain for so much of what makes us human: to reflect, to develop empathy and love, even to process difficult feelings. If we're always using technology to do everything for us, do those parts of being human start to wear away?
I don't have the answer, but it's a question I keep coming back to.
Support, not substitute
I'm not against AI in education, or anywhere else. It has real potential: for accessibility, for reducing teacher workload, for supporting students who learn differently, and as a thinking partner when used with intention and awareness.
The Consciously Digital Institute, where I trained, has developed an ethical framework for working with AI that distinguishes between AI as a substitute for thinking and AI as a support for it. It's the framework I follow in my own work.
A few ways to keep it a support:
- Think first. Spend a few minutes with a question yourself before you ask AI.
- Do the hard part yourself. Write your own first draft, or work out your own answer, then use AI to check or challenge it.
- Notice the pull. When quick prompting turns into handing over your thinking, name it. I call it a Prompt Loop.
- Keep some things AI-free. Reading, writing by hand, talking a problem through with a real person.
What schools need
In my written evidence to the House of Commons Education Committee, I recommended that young people are taught about cognitive offloading, the neuroscience of attention and the business models behind AI, as core knowledge rather than one-off talks. And that teachers get training in how these tools work and affect thinking, not just how to use them.
Teaching young people to use AI "safely and responsibly" isn't enough if they don't understand what it's doing to their thinking. What is attention literacy?
Questions people ask
What is cognitive offloading?
Cognitive offloading is handing mental work over to something outside ourselves, like a calculator, a list or AI. It’s useful, but relying on AI for our thinking itself may weaken skills like critical thinking over time.
Does using ChatGPT affect your brain?
Early research suggests it can. An MIT study found lower brain engagement when people used ChatGPT to write essays. The research is still new and small scale, and a lot depends on how AI is used.
Should students use AI?
AI can support learning, especially when it’s used to check, challenge or deepen students’ own thinking. The risk is when it replaces that thinking. Students need to understand the difference.
How can I use AI without losing my critical thinking?
Think about the problem yourself first, do the hard part yourself, and use AI to support rather than replace your thinking. Notice when quick prompting turns into handing everything over.
Read my full written evidence to the Education Committee, published by Parliament.