Using AI While Growing the Human's Intelligence

"Cognitive Surrender" or a Chance to Think More Deeply?

“Use it or lose it!” Most of us have heard that at least once. We often speak of it in terms of human talents and abilities. For instance, playing a sport. Public speaking. Playing a musical instrument. Thinking. Focusing.

Think, or lose your ability to think. Focus, or lose your ability to focus. The brain as a muscle. Exercise it, or it will atrophy.

These thoughts were sparked in me as I read the Washington Post article “How to Stop ChatGPT from Ruining How You Think” by Michael J. Coren, published in the Post’s July 7, 2026 online edition (and unfortunately behind a paywall). He begins with a question: “Can you use AI in the mental gym without losing cognitive muscle?“

Coren shared the results of studies that have shown what happens when smart humans – in this instance, Boston Consulting Group consultants, are asked to solve math problems using AI, and then are switched back to solving similar problems without help from AI. The findings? While using AI, the consultants did correctly solve more problems than a control group that did not use AI. But then after using AI for just ten minutes, when the consultants had AI taken away and were back to solving the problems on their own, the percentage of correct answers was worse than the control group.

Another interesting finding: once the AI was taken away from them, consultants gave up on trying to solve problems faster than they did before their use of AI.

The article goes on to introduce the term “cognitive surrender,” used by Wharton researchers to describe the phenomenon they observed in a separate study of 1,300 people who were asked to choose between two choices in answering math problems – an intuitive but incorrect one generated by an AI platform, or to come up with their own answer (the correct answer required slower reasoning to arrive at). Many people, seeing the AI-generated intuitive answer, chose it instead of relying on their own brain to arrive at the correct solution.

The Wharton researchers observed that the users of AI “relinquished cognitive control and adopted the AI’s judgment as their own.”

Using AI - We've Been Here Before

Does this evidence support the argument some people make that AI will become smarter than us and eventually rule over us (or do worse to us?) Not so fast! Coren points out that other points in human history have raised similar fears that advances in technology would curtail humankind’s ability to think – going all the way back to the advent of the written word. This recalled my experience in the mid-1980s as a junior Air Force officer being told to make use of a new spreadsheet tool called Lotus 1-2-3 (younger readers, substitute in Microsoft Excel if that helps!)

My objection was that it would cause me to become worse at math and accounting, because I’d no longer be using my 10-key calculator and my brain. Delegating math tasks to a machine seemed so irresponsible to me back then! But I soon learned that this new spreadsheet software actually expanded my ability to perform higher-value functions, such as variance analysis, planning, forecasting, reporting. I had more time for all this because I didn’t need to spend as much time on adding, subtracting, multiplying and dividing.

But just as with today's Artificial Intelligence revolution, yesteryear's spreadsheet revolution required that I still engage my cognition to check and guide the ghost in the machine.

Crucially, I still needed to check my spreadsheet. Did a formula get copied into the correct cell? Did something become corrupted? Human troubleshooting was still needed but in general, it became much easier and faster to spot and correct a problem.

As a result, my ability to use numbers to analyze, plan, forecast and report grew stronger, precisely because I had more time to engage in these tasks. I was no longer burning through hours of looking for a slipped decimal point, transposed numbers, or a calculation error.

I’m certainly not unique. There are most likely millions of folk alive today who remember the great spreadsheet revolution of the 1980s, part of the larger personal computing revolution, who could share similar stories.

Toward a Smart Path Forward

Back to the original question: “Can you use AI in the mental gym without losing cognitive muscle?” Coren says “The answer boils down to deciding what work we want to do and what we’re willing to off-load to machines.” His article hints at many other good practices that are no doubt already familiar to some of you:

• Prompt wisely and carefully, providing as much relevant information and instruction as possible

• For quantitative work: Give your AI tool any tedious work to keep your mind free for analysis and decision-making

• For editorial and point-of-view work, reserve initial idea generation and final drafts for yourself; use AI in the middle part of a project to add depth and perspective you may have missed

• Ask for the tool to identify its sources and check those sources out for yourself

• When using AI to help develop a point of view, ask the tool to find contrary source material and use it to argue against itself

• Inspect answers and push back/ask for further explanations when the output looks possibly suspect.

These practices help you to use your preferred AI tool as a critical thought partner instead of as a substitute for your human wisdom and insight. They keeps you in the driver’s seat, and allow you to stay in control of your AI-assisted work and not the other way around.

Schedule an Exploratory Call