The Cognitive Brain is Surrendering Critical Thinking to ChatGPT

As the world argues about the meta-level risks and benefits of AI in its various forms, why am I as a clinical and consulting psychologist thinking about the—seemingly narrow—cognitive effects of AI? 

Of course, our cognitive capacities matter in day-to-day life and are essential in our work roles. 

But for me there’s a personal dimension. 

My father went to work at MIT in the early 60’s into the 80's in order to work in robotics and AI. That’s widely considered when and where the field of AI was born. 

When my brother and I trailed along with my dad to his office at MIT, where we twirled around on the swivel chairs in rooms filled with refrigerator-sized computers, we were overhearing conversations about questions such as the possibility that computers might be able to replicate the functions of the human brain. 

My dad tried out a primitive "beta" robotic dog that we operated with a remote control in our kitchen.

Nerd out with me here for a few seconds (or skip this paragraph) as I offer my “non-tech” brief history of AI, and my dad’s role in it: My father worked on applying the concept of “Artificial Neural Networks (ANN),” artificial networks that are inspired by the structure and functions of biological neural networks in our brains. Today, ANN are central to “machine learning” and generative versions of AI, such as Claude, Gemini and ChatGPT. 

The complex human-AI relationship is in my blood. Or I should say, my brain. And my mind. 

Groundbreaking studies conducted in 2025 and 2026 at MIT and The Wharton School, among other institutions, have found strong evidence of "cognitive atrophy" when we turn over our thinking and problem-solving to AI. Reliance on AI degrades memory and executive functioning, as well as creativity, reasoning, and independent thinking. 

As our brains become dependent on AI, our ability to exert the cognitive effort to problem-solve degrades.

We are not just transferring executive functioning tasks. We are off-loading decision-making to machines. Importantly, this cognitive transfer occurs without our realizing it.

Our capacity for critical thinking is eroding. 

We are in the midst of a “cognitive surrender” of our reasoning abilities to AI. 

Performance pressure and overload are factors as well.  According to Prof. Mindy Shoss, who studies AI use in organizational settings, “If people are overloaded or feel pressure to simply produce more, AI will be used to speed up existing work. Organizations therefore need to create time and incentives for employees to explore, learn, and build on their existing strengths.” [Source: APA] I would also note that from a psychological standpoint, the increase in performance pressure adds to stress, exhaustion and burnout, and “AI anxiety,” especially among senior managers.

When we think through problems on our own first, then turn to AI to obtain feedback or supplement our own independent thought process, AI can enhance critical thinking. 

Certainly, AI tools such as ChatGPT, Claude and Gemini for search purposes can achieve results with vast efficiency and scope that surpass our human capacities, with little or no downside to our cognitive functioning.

What we’ve learned in our short time interacting with ChatGPT, Gemini and Claude is that we need to make nuanced, informed decisions about how we use these powerful new tools. If we are deliberate in our decisions- as individuals, as organizations- we can offload tasks at which AI excels and save our cognitive resources for complex problem-solving that may entail judgment, empathy or ethical reasoning.

As leaders of organizations, we need to think about when and how it most serves us as human beings to use AI. 

And we need to learn from experience. We are only a few short years into chatbots. As we learn, we will need to rethink our organizational practices for interacting with AI.

This is not an all-or-none situation. As in most matters, there are mission-critical tradeoffs to weigh.  

AI is many things. But it’s not a blunt instrument. 

For my team, we avoid using AI other than for search purposes. I used AI for this post, but for search only. I write my own emails. I wrote this post. Entirely.  

We should use our uniquely human cognitive capacities to decide when and how we want to make use of the various AI tools available. 

And—this is important—we need to avoid anthropomorphizing AI, even though chatbots use the pronoun “I.” Let’s not allow ourselves to be seduced. Or deluded. 

So consider this: Do you want to surrender your critical thinking to AI? 

P.S. Looking for an overview of the psychological research on the cognitive effects of AI and the implications for the workplace?  Look no further: See Abrams, Z. (2026, July 1). How AI is reshaping human skills and thinking. Monitor on Psychology,