AI can write, calculate, research, summarise and even generate ideas in seconds. What once required considerable time and effort can now be done almost instantly. This is undoubtedly a great advantage.

But is there a danger hidden in that convenience?

We develop our faculties by using them. Thinking, writing, questioning, analysing, making mistakes, correcting them and trying again, are not merely ways of producing an end product. They are the process through which we learn and develop our ability to think.

If we routinely ask AI to do the thinking, writing and reasoning for us, we may get the answer without acquiring the ability to arrive at the answer ourselves. This is particularly relevant to the younger generation. 

A person who has not yet developed these faculties may unknowingly use AI not as an aid to learning, but as a substitute for learning.

There is a growing body of research examining precisely this question: what happens to human cognitive skills when we increasingly rely on Artificial Intelligence?

Brooke Macnamara, PhD, an associate professor of psychological sciences at Purdue University, US who studies how we acquire skills and how they can degrade when we do not keep them sharp, puts the question succinctly:   

“There’s clear evidence that people can perform many tasks better with AI. The question is: What happens to our skills if we start relying on it regularly?”

That, to me, is the more important question about AI.

AI and Critical Thinking

Some early research suggests that certain skills may weaken when we too readily offload them to AI. Other research indicates that the relationship is more complex. Passive use can lead to skill decay, while structured and deliberate use may actually improve critical thinking and creativity.

A study of 319 knowledge workers, conducted by researchers from Microsoft and Carnegie Mellon University, found that people who had greater confidence in generative AI also reported engaging in less critical thinking when using it for work tasks. These tasks included developing ideas, learning about new topics and seeking guidance when making decisions.

Mutlu Cukurova, a professor of learning and artificial intelligence at University College London, observed:

“At the moment, most interactions between people and generative AI are transactional. You ask a question; you get an answer; you move on.”

He argues that such interactions are more likely to lead to cognitive atrophy than cognitive development.

Research led by Shiri Melumad, an associate professor at the Wharton School of the University of Pennsylvania, offers another interesting example. Participants researched subjects and then wrote advice for an imagined friend. Those using AI tools spent less time searching and produced shorter, less detailed summaries. Readers rated the AI-assisted advice as less helpful, less trustworthy and less likely to be adopted than advice produced by those who conducted their own internet research.

Melumad's observation is particularly relevant:

“People learn best when they engage in active exploration, problem-solving, and sensemaking.”

If an AI system performs that work for us, learning can shift from an active process to a passive one.

An MIT study by Kosmyna and colleagues examined brain activity during essay writing and found weaker neural connectivity among participants who used an LLM compared with those who used a search engine or no outside tools. However, that study had not yet been peer reviewed, and its findings therefore need to be treated with appropriate caution.

AI itself is not the Problem

There is another side to the research.

Michael Gerlich, a professor at the Swiss Business School in Zurich, studied 150 participants writing essays under three conditions: without AI, using ChatGPT without guidance, and using ChatGPT with structured prompts designed to encourage deeper thinking.

The structured prompts required participants to think independently, conduct targeted research, construct and critically review an argument, and revise their work.

The essays produced with structured prompting received the highest scores from expert reviewers. That is an important distinction.

AI can Replace Thinking or Stimulate Thinking. The difference lies in how we use it.

Gerlich suggests reversing the conventional approach. Instead of giving AI all the information and asking it to produce the answer, first think through the problem ourselves. Ask AI specific questions when additional information is needed. Reach our own conclusion and then give the problem and our solution to AI for criticism and feedback.

That approach keeps the human being in the driver's seat.

A Force Multiplier

There is no need to fear AI or avoid using it. Use it to extend your capabilities, save time and challenge your thinking. But do not outsource the thinking itself.

For someone who has already developed the ability to think, analyse, write, solve problems and make decisions, AI can be an extraordinary productivity multiplier. Routine work can be completed in a fraction of the time, leaving more time for things that require human judgement and experience.

For a 74-year-old who has developed these faculties over decades, AI can be a remarkable force multiplier. It compresses routine effort and potentially allows ten remaining years of life to contain productive output that might otherwise have taken much longer.

For a young person, the equation can be completely different.

The struggle to calculate, remember, write, research, argue, solve a problem and make a mistake is part of developing the brain. If AI routinely removes that struggle before those faculties have matured, the immediate gain in efficiency could come at the cost of long-term cognitive development.

A child learns mathematics by doing mathematics. A writer learns to write by writing. A student learns to research by searching, comparing and questioning. A person develops judgement by making decisions, sometimes making mistakes and learning from them.

If AI performs these developmental activities before the underlying faculties have been formed, we may create a generation that is extremely efficient at obtaining answers but increasingly dependent on a machine to formulate the questions, examine the evidence and construct the reasoning.

That is a very different problem from AI replacing human beings.

Is Banning AI the Answer?

I do not think banning AI in schools and colleges altogether is a good or practical idea. It would deny students an important technology and, more importantly, fail to teach them how to use it responsibly. But I can see a strong case for AI-free periods, assignments and examinations, particularly in foundational education.

Students should first demonstrate that they can think and perform a task themselves. AI can then be introduced as a tool to extend those capabilities. Technology has always freed humans from effort. But some effort is actually necessary for human development.

The challenge is to distinguish between unproductive effort that technology should eliminate and developmental effort that we should continue to undertake.

This applies beyond AI. Writing, arithmetic, memory, navigation, research, decision-making and eventually even judgement can all be affected when technology routinely performs the functions through which those faculties are developed.

The issue, therefore, is not whether technology should be resisted. It is whether we retain the ability to use our faculties even when technology makes it unnecessary to do so.

The real question is not whether AI will make us smarter or less intelligent.

It is: Will we continue to exercise our own faculties when a machine can do the work for us?

The frightening possibility is not that AI becomes more capable than humans. It is that humans stop exercising capabilities they already possess, or have not yet developed, because AI can perform those functions for them.

The real danger of AI may not be the loss of human capabilities to machines, but the degeneration of human faculties through disuse.

Perhaps, therefore, the right question is not:

“What can AI do for us?”

but:

“What must we continue to do ourselves so that we remain capable of doing it?”

AI should be an aid to the mind, not a substitute for the mind.

(The author is an Indian Army veteran and a contemporary affairs commentator. The views are personal. He can be reached at  kl.viswanathan@gmail.com )