Why Thinking With AI is a Brain Health Strategy

t was early. Coffee next to me. A piece of work I knew I needed to think through properly.

Not urgent. But important.

I opened AI instead.

Typed a quick prompt. Watched it come back structured, clear, usable.

For a moment, it felt like progress.

It wasn’t.

The risk isn’t that AI gets things wrong. Most of the time it gets them right. That’s actually the problem.

It gets them right before you’ve done the work of thinking.

That’s a different kind of risk. Quieter. Harder to notice. And from everything I’ve read in Dr Tommy Wood’s work on cognitive reserve, far more consequential than most people realise.

What the brain actually needs

The research is pretty clear on this, even if the popular conversation around AI rarely touches it.

The brain builds capacity through demand. Not passive input. Not consumption. Demand. The kind of effort that requires you to wrestle with something, hold competing ideas in your head at once, push past the first answer, and arrive somewhere you hadn’t been before.

Dr Wood calls this headroom. The reserve the brain builds when it’s regularly asked to do hard things.

Most people assume AI adds to that reserve. It doesn’t. Not automatically. What you get out depends entirely on what you bring in.

AI does not improve thinking. It amplifies the thinking you arrive with.

If you arrive with a well-formed question, a point of view, and a willingness to push back the exchange is genuinely stimulating. The brain works. Headroom is built.

If you arrive with nothing, the tool fills the space. Efficient. Useful. And quietly removing the very friction that builds cognitive strength over time.

The thing I haven't been able to unsee Add Your Heading Text Here

Watching my dad lose his capacity to think clearly slowly, in patterns, over years, changes how you hold this question.

You don’t take thinking for granted once you’ve watched it disappear.

Dad was sharp for most of his life. Methodical. A man who worked things through. And then gradually, the process became effortful. Then unreliable. Then gone.

I’m not drawing a straight line between AI use and dementia. That would be irresponsible.

But I am saying that once you’ve sat with what it means to lose the capacity to think, really sat with it, you develop a different relationship with anything that makes thinking optional.

Watching my dad lose his capacity to think clearly slowly, in patterns, over years, changes how you hold this question.

You don’t take thinking for granted once you’ve watched it disappear.

Dad was sharp for most of his life. Methodical. A man who worked things through. And then gradually, the process became effortful. Then unreliable. Then gone.

I’m not drawing a straight line between AI use and dementia. That would be irresponsible.

But I am saying that once you’ve sat with what it means to lose the capacity to think, really sat with it, you develop a different relationship with anything that makes thinking optional.

A framework that is honest

I’ve been thinking about AI use in three categories. Not as rules. Just as a way of staying honest with myself.

Substitute

No thinking first. Just asking.

The output replaces your thinking, and you accept it. It feels like speed. It registers in the brain as demand removed.

This is where Thought Architecture is absent. And where the long-term cost quietly accumulates.

Assist

Partial thinking, then AI.

You have a rough sense of the problem. The output improves your thinking but doesn’t replace it. Better. Still inconsistent.

Stimulate

You think first. Then you engage.

You arrive with a point of view. A structure. A specific question worth testing. You use the response to challenge what you already think, not to fill a gap you left open.

This is where AI and brain health intersect. The cognitive demand is real. The thinking is yours. The tool sharpens it rather than substitutes for it.

The difference between these three isn’t the tool. It’s what you do before you open it.

The practice

This doesn’t require a new system. It requires a small pause which, as anyone who’s read the earlier pieces in this series will recognise, is exactly where most of the real decisions in brain health happen.

Before you open AI, spend two minutes on three questions:

What is the real problem I’m trying to solve?

What do I already think is true?

What would I need to see to change my mind?

Write them down, even roughly. Then bring that to the conversation. Push back on what comes back. Restate the answer in your own words before you use it.

That is Thought Architecture. I’ve written about it in full elsewhere if you want the complete framework, it is worth reading. But in the context of brain health, it comes down to one small habit before you open a chat window.

The prompt you send says something about the thinking you did before you sent it. Most of the time, if we’re honest, it tells us we didn’t think first.

Where this sits in the series

This series started with a reframe: the brain is not something you protect, it’s something you build.

The second piece made that personal what building improvements in your brain looks like on an ordinary Tuesday, across six daily practices.

The third piece went somewhere harder. What my dad’s journey through Lewy Body Dementia taught me before I finished the book and what it still teaches me now.

This piece adds something that the science and the personal story both point towards.

The tools we use most often are not neutral. They either place real demand on the brain or they quietly remove it. And in an environment where removal is now effortless, the choice to engage differently must be deliberate.

Thought Architecture is how you make that choice consistently.

One conversation at a time.

Next: Piece 5 — The Headroom Principle. What it means to build an organisation that doesn’t run out of thinking capacity either.

Some of the thinking in this series was sharpened by reading The Stimulated Mind by Dr Tommy Wood a book I recommend without reservation to anyone who wants to understand the science behind what is explored here.