The AI Paradox: Thinking Smarter or Surrendering Thought?

In imaginings: In ancient Athens, a philosopher ambles through the public marketplace, vigorously scaring people about the dangers of writing. Even Socrates isn’t a fan — he worries that when folks have the written word to lean on, they’ll come across as wise without actually knowing anything. Flash forward to today — a corporate strategist reads a generative AI summary of a complex report, nods, and moves on. The tools may have changed, but the question has not: Are we truly expanding our minds or simply passing the heavy lifting to machines?

A recent study, The Impact of Generative AI on Critical Thinking (Lee et al., 2025), explores an interesting paradox. The ability of AI to carry out more of the heavy lifting in knowledge work means that the way we use our brainpower is beginning to change. We are doing less deep probing now — we’re controlling machines that think for us. But what takes place when we stop questioning and begin receiving passively?

The Night Before Writing

(I was trying to finish writing)

Socrates was one of the first to denounce our proclivity to reach for intellectual shortcuts. He was not a fan of writing, believing it gave people the false sense that they knew things without actually knowing them. For him, thinking meant more than keeping information in a ledger — it meant grappling with ideas and integrating them. Fast forward to present day, and we’re in a very similar situation with AI. Would we cease to think if all the thinking was done by machines?

Lee et al.'s study suggests we might. They become dependent on AI, and their critical thinking skills erode. The corollary of Aristotle’s phronesis — practical wisdom — is that we can’t outsource intellectual depth; it has to be cultivated actively. With no friction, if we just accept the AI-generated insight, innovation will wither.

AI Hidden: On Foucault, Derrida

If Socrates feared the outsourcing of thought, Michel Foucault would argue the point was in some ways deeper — that A.I. changes not just the work we do to think but also what we think. His notion of biopower argues that technologies do not simply provide information, but that they shape the systems by which we are able to make meaning of reality. AI, in this sense, doesn’t just help us think; it quietly decides what is worth thinking about.

Derrida, the deconstructionist, might ask whether AI is grubbing its fingers where we have failed to feel. No piece of writing — or A.I. output — is without bias. Experts struggle to refine AI-generated text in fields where their own expertise is foreign to AI, forcing professionals to accept, rather than challenge, AI’s prejudices, the research found. As they define relevance, algorithms constrain human interpretive power, achieving the quietest censorship of critical dialogue. 

The AI is a partner, not a replacement.

But history reminds us that new technologies don’t only eat away at cognition — they can expand it as well. The printing press didn’t kill deep thinking; it ended a way of distributing knowledge. The same could be said of AI, but only if we wield it with care. Lee et al. do offer a few suggestions for keeping A.I. a force for augmentation, an augmentation of our own intelligence:

- Create Deeper Thought – AI should not just provide quick answers, but rather prompt minds. The best systems should challenge users to doubt, iterate, refine, to interact actively.

- AI Logic Transparency – Understanding how AI reaches conclusions helps prevent blind trust and enables critical reflection.

- Augmenting Human Capabilities – AI must complement human intelligence, not replace it. We need to remain involved with our work and not solely rely on the deductions that machine learning will produce.

The real question is: Will we harness AI to enhance our intelligence, or will we become so dependent on it that we forget how to think? If we listen to Socrates and Derrida, we’ll come to see that AI doesn’t supersede human cognition — it’s a mirror, both echoing our capacity and exposing our vulnerabilities. The real question isn’t whether robots will do our thinking for us. It’s whether we will allow it to change what we think is thinking.

Author: Amine Mekkaoui, CEO of ZenBee.io, and Managing Partner of Croyten is a visionary leader driving innovation in technology and inspiring the next generation of entrepreneurs.