Artificial Intelligence and Cognitive Decline: Can We Detect the Quiet Transformation of Our Thinking?
God Translation Bureau is a compilation team under 36Kr, focusing on fields such as technology, business, the workplace, and life, and mainly introducing new technologies, new ideas, and new trends from abroad.
Editor's note: Faster thinking does not equal deeper thinking. AI has significantly improved the efficiency of thinking and work, making content production more efficient and smooth, but there is a hidden risk of cognitive decline that is easily overlooked. This article is from a compilation, hoping to inspire you.
Key points:
Thinking speed can increase, but thinking depth may not increase synchronously.
External work performance continues to improve, but the underlying thinking mode is quietly changing.
Superficial results are steadily getting better, but internal cognition may deviate in another direction.
Image source: Pixabay
Previously, I proposed the concept of "thinking speed". Today, I want to further explore this idea, focusing on the current situation where artificial intelligence participates in and assists human thinking. In the past few decades, the machines we developed only simply accelerated the spread of information. However, large language models are completely different. What they change is not just the efficiency of information transmission, but the way of thinking itself.
When using AI to think, new questions will arise from existing ones, and originally unrelated creative concepts will be connected. This vivid and dynamic thinking experience is something I've never had before.
But as I use it for a long time, the nature of things gradually becomes complex. I initially defined this change as "thinking acceleration", which means that thoughts seem to have the impetus to move forward. Later, I discovered two key phenomena: one is cognitive inertia, where people no longer think actively but directly hand over the thinking work to AI; the other is the pseudo - intelligence phenomenon, where the content output by AI is smooth and logical, but it doesn't understand the content it generates.
After in - depth research, I concluded that thinking can become faster but shallower; work performance can continue to improve, but human's active participation in thinking is constantly decreasing. This also makes me start to reflect: perhaps it's wrong for us to measure the change of thinking by "speed" all the time.
Artificial intelligence is reshaping the essence of intelligence
For a long time, we've believed that human intelligence is a relatively stable ability, a fixed quality polished by learning and experience. However, the emergence of artificial intelligence has completely broken this inherent rule.
The upper limit of human intelligence may not have changed, but the development trajectory of thinking has completely changed. What really matters is never how much thinking ability we have, but whether we are still actively using this ability and in what direction our thinking is developing.
Speed has never been the core standard for judging the quality of thinking. Just like a car, no matter how fast it is, if the direction is wrong, everything is in vain. The same is true for human thinking. Nowadays, AI is no longer just simply providing us with answers, but deeply integrating into our entire thinking process and subtly changing the underlying logic of thinking.
What I really care about and worry about is not whether our thinking has become faster or slower, but the relationship between humans and thinking itself is quietly changing. Are we still insisting on asking questions actively, conducting in - depth research, and relying on our unique thinking ability to precipitate cognition and leave thinking traces?
Three development trajectories of human thinking under human - machine collaboration
Human thinking has never been a straight line. Based on my long - term experience of using AI, I found that in human - machine interaction, human thinking mainly shows three completely different development directions.
1. Positive upward type
This is the thinking mode closest to benign iteration. In the dialogue and interaction with AI, we constantly generate new questions and open up new thinking paths along the dialogue. Our curiosity is constantly stimulated. We may just want to find an answer at first, but finally, we come up with more valuable and in - depth questions. Even after the interaction ends, the inertia of thinking still exists, and our thoughts will continue to extend and iterate.
2. Flat and stagnant type
This state is very hidden and difficult to detect. On the surface, we've been thinking and interacting with AI all the time, and we haven't stopped thinking, but we haven't produced any substantial results in the end. The brain seems to be in an active state, but after it's over, we can't really tell what we've thought about and what we've gained.
This is not due to laziness or distraction, because the cognitive function of the brain has been operating, but it has fallen into a state of ineffective thinking and stagnation.
3. Reverse downward type
This is the most dangerous and difficult - to - identify state because it can disguise itself as "progress". Using AI to write emails and create content, the efficiency is greatly improved, and the work process becomes extremely smooth without any blockage or hesitation. This visible high - efficiency can easily make people think that they are continuously making progress.
Thinking needs resistance, and in - depth cognition cannot be separated from the "feeling of effort"
I always believe that true in - depth thinking always requires appropriate resistance and running - in. The seemingly painful process of confusion, repeated deliberation, painstaking research, and struggling with fuzzy problems is precisely the key to shaping thinking and precipitating oneself. Even if we don't get a perfect answer in the end, this thinking process will actually improve our cognitive ability.
As AI takes over more and more mental work, a hidden cognitive gap is forming, and we can't even detect it or have tools to measure it. The external work results are getting better and more professional, but the underlying thinking logic and thinking habits are quietly deteriorating.
Simply put, we can produce higher - quality content, but we are becoming less and less able to think and create actively. The external work performance continues to upgrade, but the internal cognitive function is quietly degenerating.
In the past, we always judged the level of intelligence by "external outputs" such as exam scores and professional achievements. This way of judging was reasonable in the past because the internal process of thinking was completely invisible, and we had no way to peek into the details of the brain's thinking. We could only judge ability based on the final results.
But AI has changed all this. Nowadays, human thinking will leave clear traces: which questions we actively delved into, which questions we gave up halfway, which thinking we directly handed over to AI, and which difficult problems we chose to avoid. For the first time, we can intuitively see the real operating trajectory of human thinking.
This is also the core view of the whole article: External work performance and internal cognitive quality are no longer developing synchronously. The gap between the two is the biggest cognitive risk in the AI era. AI can optimize our thinking results, but it is quietly changing and even weakening our core thinking ability to produce results. The surface achievements are rising, but the underlying cognitive trajectory is deviating in the opposite direction.
My deepest worry is that these invisible changes in thinking will gradually solidify and even become permanently fixed without anyone noticing. By the time we really realize the problem, we've already adapted to this inefficient and shallow thinking mode. What's even scarier is that we'll completely forget the state of original in - depth thinking and can no longer detect what we've lost.
Translator: Teresa