Entrusting Spilled Thoughts to Light the Margin for Thinking: Does Generative AI Rob Us of Our Ability to Think?

A paper from MIT has been released stating that "using ChatGPT reduces critical thinking skills".

It claims that brain activity is at its lowest when writing using ChatGPT.
This feels like a strong impression of rehashing old arguments every time something new emerges.

In the past, people used to say, "Watching TV makes you stupid."
Playing games, reading manga, surfing the net, using social media...
The subject changes, but the argument is repeated.
It is nothing more than continuing to stoke the same sense of crisis.

Needless to say, if you just dump the thinking process onto generative AI and get the result, your brain activity is unlikely to increase.
On the other hand, if you use generative AI as a tool within the thinking process, your contemplation can deepen.
That is all there is to this story.

However, there is something I think about every time I hear this kind of news.
Is a "state of not thinking" in a specific task really a problem?

I don't think so.
Even before generative AI was born, people have been simplifying and automating their thinking.
Outsourcing thought is even a part of human activity.

For example, even in the mundane chore of "cooking rice," this is the case.
Originally, it involves a series of thoughts, such as the soaking time considering the temperature, the intensity of the heat, the timing of the boil-over, and the steaming time.
The complexity of those judgments should not be underestimated.

Of course, in modern times, these are left to the rice cooker, and we hardly think about them at all.

There is no shortage of similar examples.
Searching for a destination route using a car navigation system or the automatic temperature control of an air conditioner are the same.

These certainly create a "state of not thinking" in specific tasks.
Yet, it is not generally considered a problem that human thinking ability declines because of them.

Then, what separates these from generative AI?
It is surely the "hierarchy."

There is a "hierarchy" to thinking.

Examples like cooking rice are thinking for judgment based on senses, experience, and rules.
These are less likely to vary between people and are less tied to individual evaluation.
Even if such relatively low-level thinking is omitted, it is hard to feel like it is being stolen, and it is simply perceived as efficiency.

On the other hand, generative AI is involved in higher-order thinking such as creation and logical reasoning.
This is a part where there are large differences between people and which is connected to identity.

What generative AI is confronting us with is that a certain degree of what is thought to be creative thinking is not as high-level as people believe.
It is telling us that it is not unique to that person.
That shakes the flame of individuality.

In other words, people do not feel anxious when they are robbed of the opportunity to think.
They feel strong resistance when it is something they believed formed a part of themselves.
They fear for their own sense of worth, which likely leads to the proclamation that "thinking ability declines," as in this paper.

It is sometimes said that "the amount of information we come into contact with in one day is equivalent to one year in the Edo period."
This is merely a sensory metaphor, and there is no clear evidence for it, but there is no doubt that we are surrounded by a vast amount of information.

Presumably, this volume of information will continue to increase in the future.

Being flooded with information and juggling numerous tasks creates a high cognitive load.
If you are under high load and have no breathing room, you cannot think creatively.
This is because thinking requires significant energy metabolism and carries a high time cost.
It is only natural to delegate parts of our thinking to generative AI in order to reduce cognitive load.

The problem is likely becoming chronically reliant on generative AI for every task, leading to a 'state of not thinking'.

When that happens, we may be headed toward the 'AI domination' often found in dystopian science fiction.

Therefore, the question of 'selective thinking'—deciding which range of our thoughts to entrust to generative AI—is what is being tested.

Caught between increasing cognitive load and the loss of identity, we must think meta-cognitively about 'what thoughts should be entrusted.'
The importance of that judgment will only continue to grow.

And when this is achieved, we will gain the necessary space to think, lighting up a creative future.

○ Related Links (ChatGPT, Generative AI related)