The AI Problem: Are We Losing More Than We Gain?.

THE AI PROBLEM

The fact that people who shouldn't be speaking about AI are speaking about it, only highlights how vast the AI problem truly is.

The issue isn't whether AI is being used more by teachers or students, HR professionals, or anyone else. The real problem lies in who should be using it, who shouldn't, and how it should be used.

It's not about whether AI has shown its benefits or not. It's about the consequences it has had on humans, not in terms of replacing them, but in terms of how much it will 'damage' them. However, this is no longer just an AI problem.

IT'S A PROBLEM OF HUMAN RELATIONSHIPS, INCLUDING OUR RELATIONSHIP WITH OURSELVES.

Hustling on LinkedIn will eventually catch up to us.

All the world's knowledge has been available on the internet for a long time. Once upon a time, there was a joke about where one should hide a dead body. The answer was on the second page of Google. The trick to using Google was how well you understood the problem and could then synthesize it into a question for Google, so there would be no need to ever go to the second page of Google.

Today, AI, be it ChatGPT, Copilot, or any other LLM, allows you to avoid even having to formulate the question. You don't even need to know what the problem is.

AND THAT IS THE PROBLEM.

Another problem is that today we can have all LLM models as apps on our phones. The problem with phones is that they are COGNITIVELY NON-INVASIVE. They have become a part of us. Just recall how many times, even unconsciously, you pick it up and 'check' something without even realizing it. And as if that wasn't enough, we've put already cognitively non-invasive phones on our wrists in the form of smartwatches.

So what happened?

The boundaries between what's on the internet and what's in our heads have been completely erased. We had a similar situation when we studied for an exam half a day before the test. We knew everything while looking at the book, but when we got to the exam, we didn't really know much. The consequence then was that we failed. Today, we don't even get the chance to fail.

By the way, this erased boundary between internet knowledge and what's in our long-term memory is called the GOOGLE EFFECT.

So now what do we have? Little knowledge in our heads and a lot of knowledge on the internet. With no further need to ever store that knowledge in our long-term memory. Because, why would we?

An empty long-term memory means we don't have a knowledge base in our heads that we can search to draw conclusions. In other words, we lack DEPTH OF COGNITION.

So, if we don't have a knowledge base in our heads, we effectively lose the ability to think critically. Long-term memory enables us not only to store information but also to connect, understand, and use it in new contexts. Without that, we merely become information transmitters, without deeper understanding or the ability to create new ideas. Without the capacity to synthesize anything at all.

AI can quickly answer questions, synthesize information, and even suggest solutions, but what happens to our ability to evaluate those solutions? We become passive users of technology instead of active creators of knowledge. In education, this is particularly problematic. If students get used to AI answering their questions, where is the room for exploration, making mistakes, and learning from those mistakes?

The next issue, which is still not AI, is metacognition. Metacognition, or the ability to think about one's own thinking, plays a crucial role in this context. The problem AI brings is not just that it reduces the need for traditional learning, but also that it undermines the development of metacognitive abilities. Metacognition involves awareness of one's own cognitive processes – understanding what we know and what we don't know, as well as the ability to plan, monitor, and evaluate our own thought activities.

And if we don't understand what we know, and even worse, what we don't know, we have the Dunning-Kruger effect. Which confirms the statement from the beginning of the text.

People who shouldn't be talking about AI are talking about it. They should just listen, draw conclusions, and store them in their long-term memory. But no...

By the way, depth of cognition means… connecting, synthesizing, understanding, critical thinking and drawing conclusions.

Metacognition means… recognizing when we don't know something, planning how to learn it, monitoring our progress, and being able to assess how effective we are at learning or problem-solving.