Boris Walbaum on the future of the mind in the age of AI

Mr. Walbaum, you have often warned of “cognitive atrophy.” Could you identify which specific mental faculties, such as synthesis, memory or critical judgment, are most at risk of degradation as we move towards a “frictionless” interface with knowledge via AI?

It is true that recent research has shown worrying declines in student performance when AI is used without a clear pedagogical framework. I believe the main concern is simply a decline in the willingness and ability to make cognitive effort, which is the very essence of learning. Learning has always required effort: the discomfort of not understanding, the effort of overcoming difficulties. This is not only the mechanism through which knowledge becomes lasting, but also the way to train our “learning muscle”. When AI provides instant, fluid responses, students risk losing their ability to learn, just as they have lost some of their ability to concentrate with social media.

Many educators argue that “prompt engineering” is the new essential literacy. In your opinion, is learning to formulate prompts a high-level cognitive exercise that can compensate for atrophy, or is it just a shortcut that further erodes our ability to think from first principles?

There is a deep misunderstanding about the use of generative AI. LLM models, as we know them today, are literally chatter machines. This means that the fundamental skills needed to leverage AI are exactly the same as those needed to have a productive conversation with an intelligent person at your disposal. First, you need to be clear and precise in your intention, which is the essence of prompting. But it goes much further: it’s also about how you react to his answers and deepen the conversation through clarifying questions, direct challenges, cross-referencing of information, like journalists cross-referencing sources. I use AI a lot and, when used well, it is highly interactive, challenging and rewarding. But you need to make sure you stay in control of the conversation. These patterns develop strategies and biases to please you and keep you engaged; therefore, it is best to remain on your guard.

Forward College places great emphasis on Emotional Intelligence (EQ). How does the development of social and emotional skills serve as a safeguard against “automation of the mind,” and why is EQ more resilient to AI disruption than pure logical-analytical intelligence?

The “automation of the mind” is precisely what AI and digital platforms exploit: our unconscious biases, habits and emotional triggers that make us predictable and manipulable. Emotional intelligence is fundamentally about self-awareness: knowing our own patterns, recognizing when we are being influenced, being intentional about the information we consume and the technologies we interact with. It is an immune system against manipulation, whether algorithmic or human.

As for why EQ is more resilient than IQ: Quite simply, logical-analytical intelligence can be fully emulated by AI, often better than humans can and certainly faster. But human relationships cannot be replicated, only simulated. AI can simulate emotions and attachment, but it cannot participate in the complex web of mutual interests, shared vulnerability, and genuine reciprocity that defines human connection.

Consider sport: no one is interested in watching machines compete with each other, yet human athletics triggers intense emotions precisely because we know that athletes are trying really hard, they are really taking risks, they really care about the outcome. The same goes for chess or video games: people train against machines, like hitting tennis balls against a wall, but when it comes to the real game, the real emotion, they want human partners or opponents. Why? Because meaning emerges from authentic relationships, not optimized performance. EQ is resilient to AI precisely because it governs what AI cannot touch: authentic human connection.

If the traditional academic essay can now be generated in seconds, how should universities restructure their assessment models to ensure they are assessing a student’s internal cognitive process rather than their ability to handle an external algorithm?

I am sincerely alarmed by the lack of awareness regarding the scale of AI-assisted cheating occurring in higher education today. When AI is banned, the ban must be seriously enforced. Universities that fail to address this risk discrediting the value of their degrees. If employers and society lose faith that a degree certifies real competence, the entire university system faces a crisis of legitimacy.

That said, when AI is permitted, it should be fully integrated into the assessment. At Forward College we apply three principles: expectations are significantly raised, because students “empowered” by AI are expected to produce work far superior to what was previously possible; students submit an “AI diary” documenting how they used it and what strategy they adopted (which is evaluated); finally, there is always an oral component to verify that students master and know how to genuinely defend their work. This approach transforms AI from a temptation to cheat into a skill that we explicitly teach and assess.

From an economist’s perspective, do you foresee a future where “deep thinking” and human-driven critical analysis become luxury goods? Are we moving towards a social divide between a thinking elite and a majority dependent on AI-mediated cognition?

This is a profound question. I believe we will soon see the rise of “Human Made” as a brand of value, just as we have seen with artisanal products. Already today, videos that go viral often stand out precisely for their spontaneity, their unequivocally human freshness. The more AI dominates the media landscape, the more people will hunger for authentic content.

So, I would rephrase your question: more than the deep thought itself, it is the authenticity that will become a premium value, even in economic terms. And authenticity comes directly from emotional and social intelligence: being in touch with ourselves, being in a genuine relationship with others. This is what produces the reactions, the creations, the moments that make life worth living, and that people will want to consume. The real divide may not be between thinkers and non-thinkers, but between those who cultivate their humanity and those who do not.

Human capital is traditionally based on cognitive resilience. If we outsource synthesis and memory to AI, how do you expect this to impact the long-term neuroplasticity and problem-solving ability of the next generation of global leaders?

Outsourcing memory is nothing new, Plato feared that writing would weaken it, then came the press, then Google. We no longer memorize phone numbers or driving directions, yet this hasn’t destroyed our cognitive ability: it’s shifted it.

The real disruption is in synthesis and problem-solving, but also in analysis and many other cognitive tasks. Google gave us access to the raw information, but synthesis remained a human job. AI does that work for us. The difference is between having the ingredients and receiving the cooked meal. But it is precisely through “cooking”, struggling with complexity, failing, reformulating, that the brain develops its capacity for structured thought.

This is similar to my previous point about learning how to learn. What concerns me is not memory, but more generally the productive mental effort that builds lasting understanding and the ability to learn. If a generation systematically delegates analysis and synthesis to AI, those cognitive circuits simply won’t develop in the same way. If these skills become rarer, they will also be more rewarded in salaries and leadership positions. There is no way around these skills when making highly complex decisions in human, organizational or social contexts, even in a world dominated by AI.

For a student starting their journey at Forward College today, what is the single “cognitive habit” or daily practice you would recommend to stay clear and avoid the trap of intellectual passivity in the age of generative AI?
Introduce yourself to every interaction, whether with the AI, a professor, or a classmate, with a question that matters to you. Not a question designed to sound intelligent, but genuine curiosity about something you don’t understand. It sounds simple, but it is extraordinarily difficult. It requires honesty about one’s ignorance and the courage to expose it. This is the essence of authenticity. And connecting it to learning is essential.

Our pedagogical model reinforces this aspect on a daily basis. In classes of 15 students based on debate and interaction, you can’t hide. Either you have engaged with the material and formed genuine questions, or you get caught. The habit of arriving with authentic curiosity rather than passively waiting to receive knowledge is what protects against intellectual passivity, whether in front of an AI or in a classroom.