I think the conclusion is a confirmation of what I had realized:
We are, in fact, at a crossroads. The first instinct — especially among those who did not live through the paper-to-digital transition — is to delegate knowledge entirely to AI: to turn to it when necessary and accept the first answer, whatever it may be. This becomes a disposable, as-needed kind of knowledge.
But the quality of AI’s answers depends on the depth of the questions. In the end, it becomes a dialogue with oneself — a way to draw answers out of our ability to ask the right questions.
I've done quite a bit with LLMs now, from legal to technical to financial topics. Without basic knowledge, the output would often have been rubbish because I kept having to ask, “Are you sure?” Without this intuition, i.e., knowledge, I'd say good night to the current capabilities of ChatGTPs, etc.
PS: It starts with the simplest word. The summary at the end contains the word “trade-off.” But how to interpret this can be completely overturned by an LLM. For me, this has a negative connotation. In German, a compromise sounds good...
Perhaps something similar has happened with keyboards versus handwriting; physically forming letters engages motor skills that aid memory, while typing just isn’t as memorable. That’s one of the reasons I still use a paper notebook in management meetings ... well, that and the fact that typing like a courtroom stenographer while avoiding eye contact feels incredibly rude.
Ancient orators used the method of loci https://en.wikipedia.org/wiki/Method_of_loci , walking as they recited to anchor their speeches in space and motion. I noticed the other day that when I re-wound and re-listened to a podcast segment, it brought me back to exactly where I had been walking the dogs a few hours earlier! I think all this is related; writing by hand or walking while reciting engages more of your sensorimotor system, which strengthens memory encoding. Embodied cognition, cross-modal reinforcement, and the ways we offload cognitive tasks to tools like keyboards or AI all shape how our brains encode and recall information.
Just as keyboards, calculators, and ChatGPT shift mental effort to tools, they also change how we learn. Offloading frees working memory but might reduce certain skills, such as fine motor memory or memorization techniques, unless consciously practiced. The method of loci works partly because it ties verbal information to spatial/motor memory. Similarly, handwriting connects visual, motor, and semantic processing. So these “old school” methods and modern tools may just be different ways of scaffolding memory. Optimistically, using tools such as ChatGPT doesn’t “dull” the brain, but just changes which neural pathways are strengthened.
Or maybe I’m just trying to remain positive and delude myself into thinking we’re not accelerating our societal slide! I think the band Missing Persons might have said it best: “It’s like the feeling at the end of the page, when you realize you don’t know what you just read.” https://youtu.be/BX86s8fHgkI
Actually, I recall an article years ago claiming that human thumbs were lengthening due to game controllers and mobile keyboard typing!
I think that turned out to be malarkey, but the way we use our thumbs has certainly evolved due more to functional adaptation and neural plasticity than anything to do with physical changes in thumb length. It's likely that both our thumb usage patterns and our brain's representation of thumb movement will continue to adapt https://youtu.be/LXzJR7K0wK0 .
Our ability to think, write, or even read beyond short answers will be on the chopping block next. I wrote a comment (https://tinyurl.com/y5n9y36u) about this under a post, but this is just a trend:
Our dependency on tools and technology has been building for centuries, but the pace of this evolution has accelerated dramatically in the last 50 years, reshaping how we think and interact with the world in unprecedented ways.
It started with innovations like writing and then the printing press, where we outsourced memory to written words. This revolutionized how knowledge was preserved and shared, and reduced our reliance on oral traditions and memory retention. In the last 100 years, radio and television have turned us into passive receivers of information. While they brought real-time news and entertainment to the masses, they also shifted us away from active participation in learning and critical thinking.
Later, tools like the calculator emerged, making arithmetic effortless and allowing us to focus on complex problem-solving. However, many have lost the ability to perform even basic mental math over time. As you said, GPS transformed navigation, enabling us to find routes anywhere easily, but at the cost of our spatial awareness and ability to navigate intuitively.
The next frontier was Google Search, which put vast information at our fingertips. This marked a shift from pondering or reflecting on problems to searching for instant answers—even on topics we might already understand. It encouraged speed over depth and made us more likely to outsource our critical thinking.
Now, with AI, we're reaching a stage where we risk outsourcing not just skills but the processes of decision-making, creativity, and ethical judgment. AI doesn't just provide information; it can generate ideas, solve problems, and even create art from patterns in the training data. While incredibly powerful, this level of dependency could lead to intellectual complacency or a loss of independent thought.
The trend is clear: with every technological leap, we've offloaded specific human abilities—memory, critical thinking, spatial awareness, and creativity—to tools. These innovations empower us initially, but they can lead to dependency and cognitive atrophy over time.
The research points to people not being able to remember what they wrote in the essay. You imply that this offloading of brain activity has negative effect. How? Only if these participants had not been able to write as good an essay the second time around then yes. Has this research confirmed this? Pls do share your thoughts.
I disagree. The very fact that people couldn't remember what they wrote in the essay is a bad thing already. It shows that no thought has gone into the essay to begin with and the people just copied from the AI. This includes all the hallucinations of AI and this is why there have been legal cases dismissed in courts in the US because law firms cited nonexistent previous cases and the judge checked and found out.
The act of writing an essay is an important way to accumulate knowledge and get better at both writing an essay and solving problems. If we mindlessly copy from AI, this learning and gathering of experience and knowledge stops and we remain ignorant.
1. AI hallucinates is a fact. So is the detrimental effect of mindlessly copying from AI. Though, this copying existed even before AI too but it became pretty much effortless now! There is no debate on these.
2. However, the mere fact that one cannot remember stuff, cannot be used as proof that AI use is negative! My point is that AI brings the good as well as the bad! Any technology has this effect.
3. All this depends on the quality of essays written using AI and without AI and if they differ significantly. Should they differ, then yes we can say people mindlessly copied and hence abused AI. Not otherwise. I find through personal experience that mindful and effortful use of AI does bring benefits. The maxim "Trust but verify" is indeed true :-)
4. In summary, I'd say that "IT doesn't matter" and hence AI too wouldn't matter once commoditized! So if we use it well, we will gain.
I am 100% with you that AI and tech, in general, have both upsides and downsides. I am also amen to the statement that “if we use it well, we will gain.”
My issue is that AI makes it easier to use technology poorly, and in my view, the experiment shows that AI is not just an evolution of Google search. It seems to be qualitatively different, both on the upside and the downside (full disclosure: I am a fan of AI and invest in the theme).
And in recent months it seems to me that the hype has become too much and we are increasingly finding out what kind of downsides AI has.
I think the conclusion is a confirmation of what I had realized:
We are, in fact, at a crossroads. The first instinct — especially among those who did not live through the paper-to-digital transition — is to delegate knowledge entirely to AI: to turn to it when necessary and accept the first answer, whatever it may be. This becomes a disposable, as-needed kind of knowledge.
But the quality of AI’s answers depends on the depth of the questions. In the end, it becomes a dialogue with oneself — a way to draw answers out of our ability to ask the right questions.
Very true and well observed. You get out of AI what you put in, so to say.
I've done quite a bit with LLMs now, from legal to technical to financial topics. Without basic knowledge, the output would often have been rubbish because I kept having to ask, “Are you sure?” Without this intuition, i.e., knowledge, I'd say good night to the current capabilities of ChatGTPs, etc.
PS: It starts with the simplest word. The summary at the end contains the word “trade-off.” But how to interpret this can be completely overturned by an LLM. For me, this has a negative connotation. In German, a compromise sounds good...
Perhaps something similar has happened with keyboards versus handwriting; physically forming letters engages motor skills that aid memory, while typing just isn’t as memorable. That’s one of the reasons I still use a paper notebook in management meetings ... well, that and the fact that typing like a courtroom stenographer while avoiding eye contact feels incredibly rude.
Ancient orators used the method of loci https://en.wikipedia.org/wiki/Method_of_loci , walking as they recited to anchor their speeches in space and motion. I noticed the other day that when I re-wound and re-listened to a podcast segment, it brought me back to exactly where I had been walking the dogs a few hours earlier! I think all this is related; writing by hand or walking while reciting engages more of your sensorimotor system, which strengthens memory encoding. Embodied cognition, cross-modal reinforcement, and the ways we offload cognitive tasks to tools like keyboards or AI all shape how our brains encode and recall information.
Just as keyboards, calculators, and ChatGPT shift mental effort to tools, they also change how we learn. Offloading frees working memory but might reduce certain skills, such as fine motor memory or memorization techniques, unless consciously practiced. The method of loci works partly because it ties verbal information to spatial/motor memory. Similarly, handwriting connects visual, motor, and semantic processing. So these “old school” methods and modern tools may just be different ways of scaffolding memory. Optimistically, using tools such as ChatGPT doesn’t “dull” the brain, but just changes which neural pathways are strengthened.
Or maybe I’m just trying to remain positive and delude myself into thinking we’re not accelerating our societal slide! I think the band Missing Persons might have said it best: “It’s like the feeling at the end of the page, when you realize you don’t know what you just read.” https://youtu.be/BX86s8fHgkI
It has. You can trace that and the rise of video games in the muscle mass in different fingers (e.g. the thumb)
Actually, I recall an article years ago claiming that human thumbs were lengthening due to game controllers and mobile keyboard typing!
I think that turned out to be malarkey, but the way we use our thumbs has certainly evolved due more to functional adaptation and neural plasticity than anything to do with physical changes in thumb length. It's likely that both our thumb usage patterns and our brain's representation of thumb movement will continue to adapt https://youtu.be/LXzJR7K0wK0 .
Our ability to think, write, or even read beyond short answers will be on the chopping block next. I wrote a comment (https://tinyurl.com/y5n9y36u) about this under a post, but this is just a trend:
Our dependency on tools and technology has been building for centuries, but the pace of this evolution has accelerated dramatically in the last 50 years, reshaping how we think and interact with the world in unprecedented ways.
It started with innovations like writing and then the printing press, where we outsourced memory to written words. This revolutionized how knowledge was preserved and shared, and reduced our reliance on oral traditions and memory retention. In the last 100 years, radio and television have turned us into passive receivers of information. While they brought real-time news and entertainment to the masses, they also shifted us away from active participation in learning and critical thinking.
Later, tools like the calculator emerged, making arithmetic effortless and allowing us to focus on complex problem-solving. However, many have lost the ability to perform even basic mental math over time. As you said, GPS transformed navigation, enabling us to find routes anywhere easily, but at the cost of our spatial awareness and ability to navigate intuitively.
The next frontier was Google Search, which put vast information at our fingertips. This marked a shift from pondering or reflecting on problems to searching for instant answers—even on topics we might already understand. It encouraged speed over depth and made us more likely to outsource our critical thinking.
Now, with AI, we're reaching a stage where we risk outsourcing not just skills but the processes of decision-making, creativity, and ethical judgment. AI doesn't just provide information; it can generate ideas, solve problems, and even create art from patterns in the training data. While incredibly powerful, this level of dependency could lead to intellectual complacency or a loss of independent thought.
The trend is clear: with every technological leap, we've offloaded specific human abilities—memory, critical thinking, spatial awareness, and creativity—to tools. These innovations empower us initially, but they can lead to dependency and cognitive atrophy over time.
The research points to people not being able to remember what they wrote in the essay. You imply that this offloading of brain activity has negative effect. How? Only if these participants had not been able to write as good an essay the second time around then yes. Has this research confirmed this? Pls do share your thoughts.
I disagree. The very fact that people couldn't remember what they wrote in the essay is a bad thing already. It shows that no thought has gone into the essay to begin with and the people just copied from the AI. This includes all the hallucinations of AI and this is why there have been legal cases dismissed in courts in the US because law firms cited nonexistent previous cases and the judge checked and found out.
The act of writing an essay is an important way to accumulate knowledge and get better at both writing an essay and solving problems. If we mindlessly copy from AI, this learning and gathering of experience and knowledge stops and we remain ignorant.
Thanks for the nice reply!
1. AI hallucinates is a fact. So is the detrimental effect of mindlessly copying from AI. Though, this copying existed even before AI too but it became pretty much effortless now! There is no debate on these.
2. However, the mere fact that one cannot remember stuff, cannot be used as proof that AI use is negative! My point is that AI brings the good as well as the bad! Any technology has this effect.
3. All this depends on the quality of essays written using AI and without AI and if they differ significantly. Should they differ, then yes we can say people mindlessly copied and hence abused AI. Not otherwise. I find through personal experience that mindful and effortful use of AI does bring benefits. The maxim "Trust but verify" is indeed true :-)
4. In summary, I'd say that "IT doesn't matter" and hence AI too wouldn't matter once commoditized! So if we use it well, we will gain.
I am 100% with you that AI and tech, in general, have both upsides and downsides. I am also amen to the statement that “if we use it well, we will gain.”
My issue is that AI makes it easier to use technology poorly, and in my view, the experiment shows that AI is not just an evolution of Google search. It seems to be qualitatively different, both on the upside and the downside (full disclosure: I am a fan of AI and invest in the theme).
And in recent months it seems to me that the hype has become too much and we are increasingly finding out what kind of downsides AI has.