🧠 Our Greatest Asset Is at Risk

Sign Up | AI Toolkit | Meet The Team

In today’s Sunday Special:

Read Time: 7 minutes

📜THE PRELUDE

Imagine you’ve just come home from work, exhausted after replying to emails, analyzing industry reports, and preparing for tomorrow’s leadership meeting.

As you collapse onto the couch and let the silence settle, you suddenly ask yourself: “What did I actually think about today?”

Not what you read, sent, or approved, but what you truly thought. An idea you explored, a sentence you crafted, a problem you solved.

You realize you can’t recall anything because you didn’t really think. Your AI Agent did most of it. What you’re experiencing is a new paradox called “Cognitive Amnesia.”

You haven’t forgotten the thoughts; you’ve forgotten the experience of forming them. The machine did the thinking. You just reviewed, approved, and moved on.

The more effortless thoughts become, the easier it is to lose track of where your mind ends and the machine begins. We’re not just automating work. We’re automating cognition.

So, when LLMs handle the heavy cognitive lifting, do we still form genuine memories, beliefs, or ideas?

💭THE EXTENDED MIND

To understand what’s happening to us, we need to first understand how our minds operate.

In 1998, British philosopher Andy Clark, along with Australian philosopher and cognitive scientist David Chalmers, proposed the Extended Mind Thesis: our minds aren’t solely confined to our brains and bodies, but can extend into the external world. So, what the heck does that mean?!

To illustrate this, they created the Inga-Otto Thought Experiment. A young woman named Inga and an Alzheimer’s patient named Otto both wanted to go to The Museum of Modern Art (MoMA) in New York City, NY.

Inga relied on her memory to recall the address, while Otto relied on his notebook to retrieve the address. Before Inga consulted her memory or Otto turned to his notebook, neither of them had MoMA’s location consciously in mind, but both would say they knew where it was if asked.

Essentially, Otto’s notebook plays the same functional role as Inga’s memory; it stores and retrieves knowledge as part of his active cognition. In that sense, Otto’s mind isn’t confined to his skull. It extends onto the notebook.

We’ve always extended our minds through external tools. For example, we often rely on a calculator to solve complex mathematical problems that are difficult to compute mentally. But when we use a calculator, we still need to understand the complex mathematical problem, decide which mathematical operations to perform, and interpret the results of the calculation.

But in recent decades, more advanced external tools like smartphones have led to cognitive scaffolding: integrating so seamlessly into our everyday lives that they function as an extension of our minds. For example, when we rely on Google Maps via a smartphone, we no longer construct mental maps of our surroundings. Instead, we follow instructions and directions passively, freeing up our attention for other things, like listening to a podcast or thinking about dinner plans.

Just like we tell Google Maps our desired destination, we ask LLMs for a desired output. The difference lies in the scope of automation. LLMs now handle the ideation, composition, refinement, and publishing of language. We’re no longer just extending our minds. Now, we’re beginning to outsource thinking itself.

🤖FROM SUPPORT TO SUBSTITUTE

When we ask ChatGPT to reply to an email, we offload the responsibility of tone, phrasing, persuasion, social context, and emotional nuance.

Before LLMs, we’d consider our relationship with the recipient, explore different ways to express our thoughts, and make definitive choices about tone, emphasis, and intent.

When we ask ChatGPT to summarize an industry report, we offload the burden of comprehension, analysis, and synthesis.

Before LLMs, we’d sift through the industry report, identify the main arguments, gather the most compelling evidence, and craft a comprehensive overview.

I’m all for the efficiency and convenience that LLMs provide, but the effects of that efficiency and convenience are worrying. They risk dulling our critical thinking by automating the cognitive processes that shape how we understand and engage with information.

This dilemma is referred to as the Efficiency-Accountability Tradeoff, which posits that the more we rely on LLMs to streamline daily tasks, the less capable we become of performing those daily tasks ourselves.

The MIT Media Lab recently published “Your Brain on ChatGPT,” which explores the mental trade-offs of using conversational chatbots.

54 participants were assigned to three groups:

They utilized ElectroEncephaloGraphy (EEG) to record each participant’s brain activity levels to assess their cognitive engagement when researching, writing, and reviewing essays.

83.3% of Group #1 couldn’t quote a single phrase from any of the essays they wrote just minutes earlier. Even scarier, 100% of Group #1 had no idea their spatial memory was weakened. For example, Group #1 experienced a decline in their ability to remember where specific sentences or certain sections were located.

What’s even more interesting is that when participants in Group #1 were reassigned to Group #3, they exhibited reduced Alpha levels and lowered Beta waves. Alpha levels are associated with calm alertness, while Beta waves are associated with strong focus.

In other words, after previously relying on ChatGPT, the brain activity levels of these reassigned participants showed reduced levels of alertness and lowered signs of focus.

The MIT Media Lab referred to this as “cognitive debt,” which essentially means you’re borrowing against future cognitive capacity for short-term convenience.

🐜THE SHRINKING SELF

When we offload cognition at scale, we learn less and become less. The boundary of self (i.e., the sense of “I thought this!”) starts to dissolve.

The more we rely on LLMs to think for us, the less we develop the capacity to think critically, reason independently, and form original judgments.

The erosion of these capacities gradually happens over time. We’ll still feel productive and produce results, but the source of those results (i.e., our own thinking) will diminish without us noticing.

This erosion leads to a sort of cognitive hollowing-out. When we don’t wrestle with language, concepts, or ambiguity ourselves, we stop forming the personal relationships to ideas that give them meaning. We lose both cognitive skills and intellectual ownership. We become vessels for information we didn’t truly absorb, echoing arguments we never constructed. The sense of “I believe this because I understand it” erodes.

Over time, the effects of this aren’t only individual but also collective. Culture evolves through thinking, debating, imagining, and grappling with profound ideas. If we let LLMs do too much of that for us, we risk becoming a civilization that’s technically powerful yet intellectually hollow.

🔑KEY TAKEAWAY

Our ability to think critically is crucial to making the right choices in life. It comes from the Greek word Κριτικός (i.e., Kritikos), which means “able to judge or discern.” If you frequently rely on ChatGPT to generate ideas or summarize texts, you risk shifting from an active thinker to a passive information consumer. We must be mindful of how we integrate LLMs into our mental processes, ensuring they enhance rather than replace our critical engagement.

📒FINAL NOTE

{{rp_personalized_text}}

Share your unique referral link: {{rp_refer_url}}