The Hidden Cost of AI That No One Talks About
Why your thinking skills are quietly eroding—and what to do about it
We’re living through the most rapid technological shift in human history. AI tools promise to make us more productive, more creative, more capable. But there’s a hidden cost that most of us aren’t paying attention to: the gradual erosion of our cognitive abilities.
The Seductive Trap of Outsourcing our Minds
I started using AI extensively during my sabbatical, playing around with every tool I could imagine—from writing songs to making videos, developing film scripts and building applications. The initial rush was intoxicating. Ideas flowed faster—so much faster. Projects materialized out of thin air. I felt like I had superpowers.
But something wasn’t quite right. It didn’t feel real. None of the outputs seemed to matter to me—and along the way I could sense a fogginess and lethargy in my brain that I haven’t known before. It was as if my ability to wrestle with complex problems, to think through things and plan for myself had been diminished. I grew increasingly dependent on external validation of my ideas, needing to run absolutely everything past an LLM: My shopping list, my communication with family members, my travel planning.
🚨 Regular GenAI Use Leads to Cognitive Decline
What I was experiencing wasn’t unique to me. An MIT study recently showed that routine use of generative AI actually contributes to cognitive decline and a reduction in logical reasoning faculties. The researchers found that participants who relied exclusively on ChatGPT showed significantly weaker brain connectivity, lower memory retention, and reduced ownership of their work. Even more concerning, these cognitive effects persisted after participants stopped using AI tools (Kosmyna et al., 2025).
This intuition was later confirmed by a groundbreaking MIT study titled “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task” that showed routine use of generative AI actually contributes to cognitive decline and a reduction in logical reasoning faculties (Kosmyna et al., 2025).
The very tools designed to augment our intelligence were, paradoxically, making us less intelligent.
👩🏽💻 Coders Forget How to Think
Similarly, a quick scroll on the AI feeds of X show many engineering leaders are talking about their teams becoming dumber as many programmers have begun outsourcing a vast majority of their jobs to Cursor, Claude Code and Codex. While the initial advantage of building with insane speed is clear, the hidden costs of losing our ability to think cannot be understated.
🧠 AI and Brain-to-computer Interfaces
With Neuralink’s “Blindsight” is a brain-to-computer interface that helps blind people see is yet another milestone in the progressive merger of human and machine. If the social media addicted smartphones in our pocket and superpowered LLMs in front of our eyes are anything to consider, we need to be careful about the secondary consequences of how we adopt and use technology.
The Agency Argument
These aren’t arguments against AI, by no means. It’s an argument for being more conscious and intentional about how we engage with technology to ensure that we are maintaining the cognitive faculties that make us human.
The question isn’t whether or not we should use these powerful tools, but instead how to use them without losing ourselves in the process.
In a world where AI can generate our thoughts, write our emails, build our applications and solve a majority of our problems for us, what does it mean to be human?
Deep focus, critical thinking, clear communication and creative agency aren’t just nice-to-haves anymore—they’re survival skills in the age of AI and automation.
I developed a set of rules to define how I work with AI to ensure that I am using it as a co-pilot to enhance my learning and cognition, not replace it. As we know from Dr. BJ Fogg’s work at Stanford, this is a tricky and slippery slope as using AI for everything is by far the easiest decision, and therefore the most likely. As such, we need to factor in the amount of motivation required to maintain agency and autonomy amidst increasingly powerful and intelligence machines.
📝 Five Rules of Responsible AI Usage
1. Own Your Genesis
Always write, sketch or voice note the first expression yourself
Don’t accept any AI input until you’ve fully externalized your initial thinking
The genesis of an idea is where your unique perspective lives. When you immediately turn to AI for the first draft, the first sketch, or the first solution, you’re essentially asking the machine to think for you from the very beginning. This robs you of the most crucial part of the creative process: the messy, uncertain, deeply human work of wrestling with a blank page.
Start every project by externalizing your own thinking first. Write a rough outline. Sketch your initial concept. Record a voice memo of your thoughts. This isn’t about creating something perfect—it’s about establishing your intellectual baseline before inviting AI into the conversation.
2. Type Every Word
Never copy-paste AI suggestions. Retype them if you decide to use them.
Force your brain to process and choose each and every element.
Copy-pasting is cognitive bypassing. When you highlight AI-generated text and paste it directly into your work, you’re training your brain to accept ideas without processing them. You’re essentially becoming a manager of machine output rather than a creator in your own right.
Retyping forces engagement. As your fingers move across the keyboard, your brain has to process each word, evaluate each phrase, and make micro-decisions about what to include or modify. This seemingly small act transforms you from a passive consumer of AI content into an active collaborator. It’s like a form of taking notes and will improve your writing or coding abilities as you internalize the suggestions.
3. Close the Tool and Step Outside
After engaging with AI, close the window and synthesize on your own
Activate your diffuse mode thinking which processes patterns
Our brains have two primary modes of thinking: focused mode (deliberate, concentrated attention) and diffuse mode (relaxed, pattern-recognizing background processing). AI tools keep us locked in focused mode, constantly consuming and reacting to new information.
Real insight often comes during diffuse mode—while walking, showering, or simply sitting in silence (Buckner et. al, 2008). After an AI session, intentionally create space for this kind of processing. Close the laptop, step outside, and let your unconscious mind work on the patterns and connections you’ve been exploring.
4. Own Your Taste
Make aesthetic, strategic and tactical decisions yourself
Use AI to generate options not make decisions
AI is excellent at generating options but terrible at developing taste. It can show you a hundred different color palettes, writing styles, or strategic approaches, but it cannot develop the aesthetic judgment that makes work uniquely yours.
Use AI as a research assistant and option generator, but reserve all final decisions for yourself. Let it expand your possibility space, but don’t let it choose which possibilities to pursue. Your taste—your ability to discern quality, beauty, and appropriateness—is one of the few things that remains distinctly human.
5. Solve Small Problems Solo
Only use AI for complex challenges or ideation
Handle routine decisions and basic problems on your own
Not every problem requires AI assistance. When you reach for AI to help with basic calculations, simple writing tasks, or routine decisions, you’re weakening your cognitive muscles. It’s like using a wheelchair when you’re perfectly capable of walking.
Reserve AI for genuinely complex challenges where its pattern recognition and vast knowledge base add real value. For everything else—the daily micro-decisions, the simple problems, the routine cognitive tasks—do the work yourself. Your brain needs this exercise to stay sharp.
🛣️ The Path Forward
These principles aren’t about rejecting AI or returning to some pre-digital past. They’re about maintaining our humanity while embracing the most powerful tools of all time. They’re about ensuring that AI serves us rather than us becoming slaves to the machine.
The future belongs not to those who can prompt AI most effectively, but to those who can maintain their cognitive sovereignty while leveraging artificial intelligence. In a world of infinite generated content, the scarcest resource isn’t information or even attention—it’s authentic human judgment, deep moral conviction, and radical compassion.
The choice is ours: we can become passive consumers of machine intelligence, or we can remain active creators who use these tools to amplify our uniquely human capabilities. The cognitive habits we develop now will determine which path we take.
Deep focus and agency aren’t just valuable in the age of AI—they’re essential. The question is: will you choose to cultivate them or give away your agency because it’s easy?
References
Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X. H., Beresnitzky, A. V., ... & Maes, P. (2025). Your brain on chatgpt: Accumulation of cognitive debt when using an ai assistant for essay writing task. arXiv preprint arXiv:2506.08872.
Buckner, R. L., Andrews‐Hanna, J. R., & Schacter, D. L. (2008). The brain’s default network: anatomy, function, and relevance to disease. Annals of the New York Academy of Sciences, 1124(1), 1-38.
Raichle, M. E. (2015). The brain’s default mode network. Annual Review of Neuroscience, 38, 433-447.
Raichle, M. E., MacLeod, A. M., Snyder, A. Z., Powers, W. J., Gusnard, D. A., & Shulman, G. L. (2001). A default mode of brain function. Proceedings of the National Academy of Sciences, 98(2), 676-682.