“How do we think?” Although it seems like a simple question, it’s actually one of the most profound a person can ask themselves.

Education can help answer this question, not through prefabricated solutions, but rather by teaching people to think for themselves, to cultivate their autonomy, and to become reflective beings capable of directing their own learning. Thinking is not memorizing: it requires methods that sharpen judgment and broaden our perspective.

• The Socratic method , for example, requires you to justify your proposals with reasonable and convincing arguments.
• The role-playing technique invites you to put yourself in someone else’s shoes, to embrace their values ​​and interests in order to understand how they make decisions.
• Comparative study , for its part, recognizes that culture and language shape knowledge: people from different geographies can reason very differently, and recognizing those differences enriches our own intellectual lives.

Taken together, these methodologies show that there are multiple ways of thinking, each of which shapes the way we deliberate, decide, and act.

In cognitive psychology, the analysis of this process is called metacognition . Practicing metacognition means becoming aware of our mental processes, detecting the patterns that guide them, and taking a step back to ask ourselves:

• What problem am I trying to solve?
• How am I approaching it?
• Am I accepting this answer too quickly?

In education, metacognition is what allows students to plan their strategies, monitor their learning, and adjust their methods. But these skills don’t emerge spontaneously; they must be deliberately cultivated.

Intuition versus analysis

A decisive breakthrough in the understanding of metacognition came from Daniel Kahneman , Nobel Prize winner in Economics in 2002, who synthesized decades of psychological research in his book Thinking, Fast and Slow . Kahneman showed that the mind does not operate as a single rational engine, but through two complementary systems.

1. System 1 , fast, intuitive, and automatic, allows you to instantly recognize a friend’s face, finish another person’s sentence, or brake suddenly before the danger even registers consciously. It’s effective, but prone to error, because it relies on mental shortcuts built from experience .
2. System 2 , on the other hand, is slow, deliberate, and analytical, and is activated when we solve a complex problem, make a detailed assessment, or question our first impressions. It’s more precise, but also more demanding, which is why we tend to avoid it.

Kahneman and his co-author Amos Tversky showed that many of our errors in judgment come from over-relying on System 1 and not activating System 2 when accuracy is crucial.

In this sense, Kahneman extends a philosophical tradition that extends from Plato and Aristotle to Hume , Locke and Wittgenstein : the search for conceptual frameworks to understand how we reason, decide and act.

Feeding the experience

The arrival of artificial intelligence has altered the balance. If problem-solving is delegated too much to machines, the risk is not only that System 2 will be underutilized, but that System 1 will also be weakened. Intuition doesn’t emerge from nowhere: it is trained and nurtured through the patient work of reason.

Every time we confront a problem, test solutions, and refine our judgments, we are feeding System 1 the experience it needs to make quick and accurate assessments in the future. If AI performs these steps for us, the store of experience shrinks. System 1 will continue to function, but its experience library will be poorer and less reliable.

The challenge is clear: if used uncritically, the tool that promises to free up our mental energy can leave us with a reduced capacity for both deep analysis and rapid intuition. And since AI not only reflects but can amplify human biases—and is also capable of generating its own hallucinations —the disciplined exercise of System II becomes more necessary than ever.

However, used judiciously, AI can help structure thinking, offer alternative perspectives, and stimulate reflection. But this requires designing learning experiences in which users compare, question, and then compare and question the machine’s results.

Instead, used carelessly, AI becomes a cognitive crutch, a way to avoid thoughtful effort and settle for easy answers.

The question is how

A recent study in China illustrates this tension well. In an academic English course, students kept reflective journals and participated in interviews. Applying experiential learning theory —according to which, learning occurs in a cycle of action, reflection, conceptualization, and experimentation—the researchers observed that students who reflected on their use of AI developed deeper and more integrated knowledge about when, why, and how to use it. Others, however, treated it as a shortcut and feared it would weaken their language skills.

The authors concluded that overuse of AI for instrumental purposes can undermine language development, underscoring the need for reflective rather than merely functional use. The lesson is clear: the decisive factor is not whether AI is used, but how.

Because?

Toyota’s Five Whys management method is a simple and powerful tool for getting to the root of a problem. The principle is straightforward: by repeatedly asking “why?” at least five times for each explanation, you leave behind superficial answers and get to the root causes. This method is a management resource but also a form of metacognition: it forces you to go beyond the first answer, to question your own assumptions, and to consciously direct your thinking.

In a world where AI can generate an answer to almost any question in seconds, the real advantage belongs to those who know how to interrogate those answers, track their own reasoning, and deliberately choose how to proceed. AI can be a partner in that process, but only if we remain the authors of our own thinking.

Therefore, it is essential to adopt practical strategies to prevent AI from eroding the essential intellectual muscle of metacognition. Some useful recommendations include:

• Define the problem before consulting AI: define the purpose, the central question, and the conclusion you hope to reach.
• Verify the sources provided by AI, select the most credible ones, and read them critically.
• Don’t let AI dictate your conclusions; anticipate them yourself.
• Work in several iterations, ideally five or more, refining the reasoning with each exchange.
• Add your own personal experiences and reflections to make the text original and unmistakable.
• When you’re done, explicitly reflect on what you’ve learned throughout the process; each reflection reinforces your metacognition.

Ultimately, metacognition isn’t about procrastination or hesitation for the sake of hesitation. It’s about taking conscious ownership of our mental processes. In an age where AI can provide an instant answer to almost any question, true intelligence belongs not to the fastest responder, but to the person who is able to question that answer, rethink it, and consciously decide how to move forward.

Author Bio: Santiago Iñiguez de Onzoño is President of IE University