Skip to content
Insights ·

Is AI making you think less? How to stay sharp: the S10 habit

By Vishal Kumar, Founder and CEO · Last updated · 8 min read

Quantana cover for an article on AI and critical thinking: a lime path with ten points rising through a purple network, representing ten solutions to one problem.

In short: AI tools make it easy to hand over thinking as well as typing. Research on "cognitive offloading" suggests heavy, uncritical use is linked with weaker critical thinking, although most of that research shows a link, not proof of cause. My answer is a habit, not a ban. Use AI every day, and keep your own mind fit. The exercise I use is S10: take one problem and come up with 10 different solutions, from your own head, before you reach for a chatbot.

I gave someone a candle on a bus.

Not a real one. We were on the coach back from a dinner at Bengaluru Tech Summit, and the conversation had turned to AI and how the next generation will exercise their minds.

So I set a challenge.

Picture your apartment building. You can go anywhere inside it, roof to basement, but you can't leave. I hand you an ordinary candle. Find 10 ways to light it.

The first answers came quickly. A lighter. Matches. The gas stove.

Then it got harder. Somewhere around five or six, the pauses got longer.

That pause is the whole point of this article.

Is AI making us think less?

Here is what I told that table:

"Cognitive decay is going to be very real with AI tools. You don't just jump to ChatGPT or Claude for the first thing. You actually have to exercise your mind to come up with solutions."

I say this as someone who uses AI all day. I run an AI company. I teach AI to working professionals and business leaders most weekends.

The worry isn't that AI is bad. The worry is the reflex.

A question appears → you open a chat window → you read the answer → you move on.

No pause. No first attempt of your own. Do that a thousand times and the muscle that generates options gets less exercise.

What is cognitive offloading?

Cognitive offloading is using something outside your head to reduce the mental work a task needs.

Psychologists Evan Risko and Sam Gilbert defined it in a 2016 review as using physical action to change the information processing a task requires, so it demands less of you. Their examples include pen and paper, GPS and smartphones. (Risko and Gilbert, 2016)

Offloading is not new, and it's not always bad. A shopping list is offloading. So is a calendar.

Their review makes one point I find useful. People tend to offload when they believe it's more efficient than using their own abilities. Those beliefs don't always match reality.

That's the trap with AI. It always feels more efficient. So we offload even the thinking we should keep.

What does the research say about AI and critical thinking?

Three recent studies are worth knowing. None of them is the final word, and I'll say where each one is limited.

1. A survey of 666 people (2025). Michael Gerlich surveyed and interviewed 666 participants. He found that more frequent AI tool use was negatively correlated with critical thinking scores, and that cognitive offloading explained much of that link. Younger participants relied on AI more and scored lower. The study is correlational, so it shows a pattern, not cause. (Gerlich, 2025, Societies)

2. Knowledge workers at work (2025). Researchers from Microsoft Research and Carnegie Mellon University surveyed 319 knowledge workers about 936 real examples of using generative AI at work. Higher confidence in the AI was associated with less critical thinking. Higher confidence in their own abilities was associated with more. They also found the nature of critical thinking shifts: towards checking information, integrating AI output and stewarding the task. The data is self-reported. (Lee et al., CHI 2025)

3. Brain activity while writing essays (2025). A team at the MIT Media Lab asked 54 participants to write essays with an LLM, with a search engine, or with no tools. On EEG, the no-tools group showed the strongest brain connectivity and the LLM group the weakest. LLM users also struggled to quote from essays they had just written. It is a small preprint, not yet peer reviewed, and the authors say results may not generalise beyond essay writing. (Kosmyna et al., 2025, arXiv)

So what do we take from this?

Not panic. A pattern.

The people who seem to keep their edge are the ones who still trust their own thinking, and who treat AI output as something to check.

666 people
surveyed: more frequent AI tool use was linked with lower critical thinking scores (correlational)
Source: Gerlich, 2025
319 workers
knowledge workers surveyed on 936 real examples: higher confidence in AI went with less critical thinking (self-reported)
Source: Lee et al., CHI 2025
54 people
wrote essays with an LLM, a search engine or no tools: the no-tools group showed the strongest brain connectivity (preprint)
Source: Kosmyna et al., 2025

So should you stop using AI?

No.

The tempting choice

stop using AI. You would give up a tool that saves hours every week.

The better choice

use AI very practically and keep your mind very fit, at the same time. I've noticed some institutions are already thinking about a post-AI world in exactly those terms.

Think of it like a lift and the stairs. You use the lift every day. You still take the stairs on purpose, because your legs need it.

S10 is my stairs.

What is S10?

S10 is a thinking game I created. We play it at Quantana.

Here is how I described it that night:

"I give you a problem, and you have to come up with 10 solutions for it. That's it. It's a great party game, and adults and kids can play it. Any age group can play it."

That's the whole method. One problem statement. Ten different solutions.

It sounds easy. It isn't, and that's what makes it work.

"Usually one, two, three, four come fairly easily. Four, five, six start getting a little tougher, and then seven, eight, nine, ten. The beauty of this is, all the answers are in your head."

The first few answers are the obvious ones, the same ones anyone would give. The later answers are where you start actually thinking. And because it's the same pattern no matter what the problem is, you get better at it each time you play.

How do you play S10?

You need a problem, a pen and a few minutes. No app.

  1. Pick one problem statement. Keep it concrete. "Light this candle without leaving your building" works better than "be more creative".
  2. Set the boundary. Say what's in and out of bounds. In the candle game, you can go anywhere in your building but not outside it.
  3. Keep one rule. You cannot break the laws of physics. Everything else is allowed.
  4. Find 10 solutions. Write each one down as it comes.
  5. Make each one genuinely different. "Ask neighbour A, ask neighbour B, ask neighbour C" counts as one idea, not three. Asking a person for a light is one way of lighting the candle. Count ways, not people.
  6. Push through the slow middle. When you stall at five or six, keep going. That stretch is the exercise.
  7. Play it with others. It works as a party game and at work. Anyone can set the problem, and anyone can play.
  1. Problem
  2. boundary
  3. one rule
  4. 10 different answers
  5. push past six

What happens when you play it?

On the bus, the answers came in a familiar order.

The easy ones first: a lighter, matches, the gas stove. Then, with more effort, rubbing two sticks together.

Then the pauses.

I had to tighten the rule halfway through. Asking a second neighbour is not a new idea. It's the same idea with a different face. That's exactly the habit S10 trains: noticing when you're repeating yourself, and looking for an answer from a different direction.

What I tell people is what I see every time. Within half an hour you get to 10, and then you wonder why you struggled so much. The answers were in your head all along.

Try it below on a lighter problem. Flip the cards one at a time and watch how the ideas change once the obvious ones are used up.

S10 flashcards
One problemThe office coffee machine is always out of beans.Sixteen ideas are hiding behind these cards. Flip them one at a time.

You've flipped 0 of 16

Flip the cards in order and notice where the ideas stop being obvious. That is the point of S10. Illustrative example ideas.
Interactive flashcards: one light problem, the office coffee machine is always out of beans. Sixteen face-down cards, Idea 1 to Idea 16, each hold a different solution. Flip a card to read it. The first few ideas are obvious and the later ones get more creative. Shuffle and Show all buttons are provided, with a count of cards flipped.
NumberIdea
Idea 1Buy more beans.
Idea 2Ask the office manager to reorder.
Idea 3Put a reorder reminder in the shared calendar.
Idea 4Hang a sheet on the machine: whoever takes the last scoop writes it down.
Idea 5Set up a bean subscription that delivers every second Tuesday.
Idea 6Rest the hopper on a smart scale that pings the team chat when it runs low.
Idea 7Start a weekly "bean captain" rota, with a small trophy that moves desks.
Idea 8Keep a shelf of your own bags with an honesty jar for refills.
Idea 9Ask the café downstairs to sell you a bag every morning.
Idea 10Everyone brings a bag on their birthday.
Idea 11Hide one emergency bag in a labelled drawer, opened only on a real crisis.
Idea 12Count cups for two weeks, then order exactly the right amount.
Idea 13Swap in a machine with a hopper twice the size.
Idea 14Make the last scoop a game: whoever empties it picks the next playlist.
Idea 15Let the machine order for itself: a rule in the supplier app that reorders at 20% left.
Idea 16Retire the machine for a Friday coffee walk and let the bean problem disappear.

Where does AI fit in?

After you've done the thinking, not before.

When you open a chatbot with 10 of your own options already written down, you read its answer differently. You compare. You notice what it missed and what you missed. You're the judge, not the audience.

That fits the Microsoft and Carnegie Mellon finding above. Critical thinking with AI looks more like checking, integrating and stewarding. You can only do that well if you've formed your own view first.

I also do the same thing at a bigger scale. Every now and then I sit somewhere away from my laptop with only a notebook and a pen, and write whatever comes. A notebook is offloading too, but it holds my ideas without generating them for me.

Think first, then ask AI
With ten of your own options already written, you read the AI’s answer differently. You are the judge, not the audience. A simple routine from this article, shown as a diagram.
Diagram: a three-step loop. Step one, think first: write your own ten options. Step two, ask AI: give it the same problem and read its answer. Step three, compare: notice what it missed and what you missed, and make the call yourself. Then repeat on the next problem.
StepWhat you do
Think firstWrite your own ten options from your own head.
Then ask AIGive it the same problem and read what it says.
CompareWhat did it miss? What did you miss? You decide.

Try it today

Pick a problem you're working on this week. Before you open a chat window, write down 10 different ways to solve it.

Which number do you stall at?

Keep reading: AI training for teams · Everyone is a tech citizen now

FAQ

Frequently asked questions

Does using AI cause cognitive decline?

The research so far shows associations, not proof of cause. Studies link frequent, uncritical AI use with lower critical thinking scores and weaker engagement. The risk looks real enough to act on, so build habits that keep you thinking.

What is cognitive offloading in simple terms?

Using a tool or the world around you to reduce the mental work of a task. Notes, calendars, GPS and AI chatbots are all forms of offloading.

Is cognitive offloading always bad?

No. Offloading memory to a list frees attention for harder work. The problem starts when you offload the thinking itself, every time, without checking the result.

What is the S10 game?

A thinking game I created: take one problem and come up with 10 different solutions. The only rule is you can't break the laws of physics, and each solution must be genuinely different from the others.

Who can play S10?

Any age group. It works for adults and kids, as a party game, and in teams at work.

How long does a round take?

It varies by person and problem. My usual expectation is that someone who stalls at six reaches 10 within about half an hour.

Should I ban AI for my team or my kids?

I wouldn't. Teach people to think without AI and with it, and to judge the output. Do your own first pass, then ask the tool.

Sources

  1. Risko, E. F. and Gilbert, S. J. (2016). "Cognitive Offloading." Trends in Cognitive Sciences, 20(9), 676–688.
  2. Gerlich, M. (2025). "AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking." Societies, 15(1), 6.
  3. Lee, H.-P., Sarkar, A. et al. (2025). "The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers." CHI 2025.
  4. Kosmyna, N. et al. (2025). "Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task." arXiv preprint.
All insights
From ambition to action

Ready to see where AI will pay off?

Get a practical roadmap: where AI will make a real difference in your business, and what it takes to get there responsibly.

Contact Global Offices
What clients say

Built on genuine partnerships.

Portrait of Toby Carrigan
“I commend Quantana for their outstanding role in developing our software; their expertise and commitment have been invaluable. What started as business has grown into a genuine partnership.”
Toby CarriganDirector · PX4 Software Resources
Portrait of Praveen Garimella
“Whatever was on our mind was very nicely and rapidly captured through the design sessions we have had and they also have the right engineering teams to get it all to life and rapidly to life.”
Praveen GarimellaDirector · Indian School of Business (ISB)
Portrait of Suzie Wing
“Feels like almost 10 years with Quantana. They are second to none, true experts in digital platforms and AI. Brought brands to life.”
Suzie WingBrand Director · Frosty Boy Australia
Portrait of Caroline Trikey
“The live product design workshop was amazing. Seeing the design come to life in front of my eyes was incredible. The process was really enjoyable, and much better than just taking notes and waiting for results.”
Caroline TrikeyFounder · Surf The Urge
Illustrated portrait of Siva Tekumudi
“I see Quantana as a rare combination of perfection, speed, and execution.”
Siva TekumudiFounder · Facillizence
Portrait of Sunil Sarvam
“It’s been a pleasure being associated with a company that continues to grow, innovate, and deliver consistently, especially its ability to embrace innovation and ride the AI wave.”
Sunil SarvamCEO · Plural Technology
“Quantana is the best choice if you want to keep your company active digitally. Their responsiveness, understanding capability, skill and TAT are extremely great. We wholeheartedly recommend Quantana.”