All AI did here is exposed an old problem in education. Kids are expected to get everything perfectly for the first time but a lot of them don't so the class gets easier next year to keep up pass rates.
We need to restructure the system to treat failure as a signal instead of a disaster. Grades should come from hard randomized exams with unlimited retakes so one bad day won't hurt you. Homework should be optional material for self study, evaluated by teachers if you choose to do it but never forced.
Hold back students for individual classes instead of a whole grade so failing one can't ruin your social life and teachers are more willing to do it. F students will realize they have to study, start actually learning and then pass on the second time. No big deal. It happened to my friends in college, no reason they can't do it in high schools.
Discipline is a skill and it's one you have to get from experience. If you try and force kids to study when they don't want to "for their own good" you're not actually helping them. Everyone needs to find their own path. Let people fail.
If you structure learning in such a way that makes learning just a means to some end, and overindex on that end being the ultimate goal, that's what you get.
This is a pedagogical problem that AI merely exposed. Educators need to figure out How to make students choose the scenic route instead of having them optimize for the most efficient completion of a task.
School has always been about creating compliant workers, not educating people. They will double down on "performing the right things" and "morals" while the economy will continue on its K shaped path as AI gets more capable.
The kids who used AI and studied for a similar amount of time as the non-AI high performers apparently had similar (slightly higher) performance. The kids who used AI and did poorly used the AI to do the homework for them (this is what the article says).
I believe AI is basically an amplifier of bad and good. I’m cynical about the world and assume it will be used more for bad than good, but I don’t doubt some of the best people in every field will be using AI to amplify their work in good ways.
It's not so clear that AI is an amplifier.. The paper has some fascinating analysis on this topic:
"At the other end of the distribution, AI students who spend more than 65 minutes on their homework receive homework and exam scores similar to those of non-AI students, suggesting that these students do not use generative AI for homework assignments. However, this group consists entirely of students who adopted generative AI no more than Öve months. Six months after adoption, no AI student spends more than 65 minutes completing their homework (see Figure A5). This is consistent with the gradual process of learning how to use AI tools. It also suggests that AI crowds out the highest level of e§ort."
"Interestingly, in the range of 50-65 minutes, the median and the interquartile range
of exam scores of AI and non-AI students are similar. This implies that, in the range
where AI students and non-AI students have overlapping homework times, students
who spend the same amount of time completing homework on average receive similar
exam scores."
"This pattern shows that students who spend the same amount of time on homework
learn similarly, with or without generative AI. In other words, generative AI reduces
time spent learning for the majority of AI students but not learning efficiency for those who spend the same time studying as the non-AI students."
The scary thing is that 81% of AI users in this study were determined to be "outsourcing" their homework to the LLMs - and the rate increases the more exposure they had to LLMs.
The dangerous thing about amplifying is that bad people are often more willing to amplify their activities, because they don't care about the negative effects. We are seeing this now as AI companies (and large companies of all kinds) rush to secure whatever advantages they can, regardless of the negative externalities. Meanwhile people who actually care about doing the right thing get trampled.
We need to shift the incentives by adding ruinous penalties for things that are currently quite commonplace if they are done by large players. Some dude training his own AI on his own computer can scrape and train. The fine for OpenAI or Meta using a single copyrighted book without permission should be in the tens or hundreds of millions.
What we're seeing currently in our society is a "loophole inversion" where the rules have an effect mainly via their loopholes. The most profitable activity is to find loopholes and exploit them as frenetically as possible to gain as much advantage as you can before the loophole is closed, or get people hooked on the loophole so it's retroactively legalized. Entities that are big enough to do this are big enough because they have lots of money behind them. Entities doing the same kinds of things without lots of money are not really doing much harm. So the best approach is to adopt a "sliding scale" in which even tiny violations by wealthy actors result in penalties enormously greater than fairly large violations by small players.
You aren’t being cynical, you’re arguing with disingenuous entities with money on the line. We already know it’s used primarily for the negative case. Everyone who ever intended High School or College knows this.
LLMs make significant mistakes frequently and smart people have no way of judging those mistakes outside their domain expertise. They are also sycophantic and great at being an echo chamber which makes people feel smart even if they are not.
So I think the burden of proof is on you to prove that they somehow amplify intelligence, it seems highly unlikely.
Most domains have some kind of internal consistency/theory building you can do. A smart person can certainly notice inconsistencies when trying to learn something. In fact they're likely to be points of confusion that the smart person will dive into just to try to make sense of things, even if they don't suspect the LLM is at fault.
It helps smart people be barely more effective and helps dumb (more importantly, people who do not value effort, people who are lazy, people who are self absorbed) people shit out endless streams of worthless tokens that can swamp out everything.
Raising the noise floor like this only makes it that much harder to find "Smart" people, which we were already doing terrible at.
I use Claude every single day, but this is such a bad tradeoff. Maybe it will help me standup a quick fix when that is needed. Maybe it can help me dig through documentation to find relevant bits and figure out the unstated assumptions underlying it. Maybe it helps me generate test cases.
Meanwhile, my day to day life is now noise. All social media is noise. All content is noise. Slop pours onto me from all directions. Writing more test cases isn't helping me.
Am I smart? Am I dumb? I don't care, right now I'm deafened
All I've noticed is AI creating a perverse incentive to make everything as complicated and bureaucratic as possible, so only the people that know how to leverage AI to cut through it ever succeed.
I'm using Claude at work myself and am impressed with the product, but notice that this is the only reason I need to use it at all. Our product pages were shit to begin with, now they're AI-generated and somehow even worse. Our procedures are incomprehensible spaghetti with enough arbitrary context switching to give a sadistic Soviet municipal administrator an erection at the thought of watching anyone try to actually follow them.
Use AI to create inefficiencies, then use AI to bypass them. Those who can't do the latter will struggle to survive.
Moneyshot: "The results are eye-opening. After six months, pupils using ai saw their average homework score rise by 18% across all subjects. The time they took to complete each assignment fell from an average of 64 minutes to 45. But come exam time, the same students scored 20% below their classmates who had not called on ai’s help. Homework scores once predicted exam performance; now those who score highest are, perversely, more likely to do worse in exams."
Study design: "David Stromberg of Stockholm University and Victor Lei and Wu Yanhui of the University of Hong Kong set out to fill the gap. They tracked 27,000 pupils aged 12-18 in China, where ai adoption has been fast. Around 80% reported using models such as Doubao and DeepSeek; the other 20% formed the control group."
As with most training the journey is the point, not the destination.
That said, I think smart use of AI could help. It could explain concepts in a way that might help you understand better, it could probe your knowledge in a more dynamic way by tailoring questions, and so on. This requires the AI be restrained by some harness, not free to write down the answers for you.
Using the forklift as a spotter and to assist in loading weights increased gains. Then again, a human can do all those things, and provide real human connection.
I used to skip homework in high school because it was boring and cut into my personal life outside of school (including my job) but I would always do very well on tests.
Then I stopped caring about tests too.
I did the same thing through college. The only reason they passed me and I got a degree is that I built the school’s website and I built personal ecommerce sites for the head of art and his wife to sell their paintings.
I haven’t done shit since like 7th grade.
Worked at Facebook, Apple, Microsoft, same behavior there - did basically nothing for them while making thousands off my games on App Store.
I know this study is focused in China but one thing I'd like to understand is the nuance in what kind of student is likely to always reach for ai for homework help as I think (especially in the US, can't speak for other countries) the varying powers that be that can determine the quality and type of education you can receive (does your country have means to buy textbooks per student or shared, do you have funds to give students ipads to take home or no, what sets the bar on how far curriculums can go, are teacher and faculty pay and incentives simply tied to student pass rates, etc.) and those do more in shaping what "use AI responsibly" will look like and why it's so different.
All AI did here is exposed an old problem in education. Kids are expected to get everything perfectly for the first time but a lot of them don't so the class gets easier next year to keep up pass rates.
We need to restructure the system to treat failure as a signal instead of a disaster. Grades should come from hard randomized exams with unlimited retakes so one bad day won't hurt you. Homework should be optional material for self study, evaluated by teachers if you choose to do it but never forced.
Hold back students for individual classes instead of a whole grade so failing one can't ruin your social life and teachers are more willing to do it. F students will realize they have to study, start actually learning and then pass on the second time. No big deal. It happened to my friends in college, no reason they can't do it in high schools.
Discipline is a skill and it's one you have to get from experience. If you try and force kids to study when they don't want to "for their own good" you're not actually helping them. Everyone needs to find their own path. Let people fail.
If you structure learning in such a way that makes learning just a means to some end, and overindex on that end being the ultimate goal, that's what you get.
This is a pedagogical problem that AI merely exposed. Educators need to figure out How to make students choose the scenic route instead of having them optimize for the most efficient completion of a task.
School has always been about creating compliant workers, not educating people. They will double down on "performing the right things" and "morals" while the economy will continue on its K shaped path as AI gets more capable.
The kids who used AI and studied for a similar amount of time as the non-AI high performers apparently had similar (slightly higher) performance. The kids who used AI and did poorly used the AI to do the homework for them (this is what the article says).
I believe AI is basically an amplifier of bad and good. I’m cynical about the world and assume it will be used more for bad than good, but I don’t doubt some of the best people in every field will be using AI to amplify their work in good ways.
It's not so clear that AI is an amplifier.. The paper has some fascinating analysis on this topic:
"At the other end of the distribution, AI students who spend more than 65 minutes on their homework receive homework and exam scores similar to those of non-AI students, suggesting that these students do not use generative AI for homework assignments. However, this group consists entirely of students who adopted generative AI no more than Öve months. Six months after adoption, no AI student spends more than 65 minutes completing their homework (see Figure A5). This is consistent with the gradual process of learning how to use AI tools. It also suggests that AI crowds out the highest level of e§ort."
"Interestingly, in the range of 50-65 minutes, the median and the interquartile range of exam scores of AI and non-AI students are similar. This implies that, in the range where AI students and non-AI students have overlapping homework times, students who spend the same amount of time completing homework on average receive similar exam scores."
"This pattern shows that students who spend the same amount of time on homework learn similarly, with or without generative AI. In other words, generative AI reduces time spent learning for the majority of AI students but not learning efficiency for those who spend the same time studying as the non-AI students."
The scary thing is that 81% of AI users in this study were determined to be "outsourcing" their homework to the LLMs - and the rate increases the more exposure they had to LLMs.
The "slightly higher" performance is based on statistically insignificant samples (between 4 and 20 students, depending on the context, out of the total population of 26,000): https://bsky.app/profile/benjaminjriley.bsky.social/post/3mt...
The dangerous thing about amplifying is that bad people are often more willing to amplify their activities, because they don't care about the negative effects. We are seeing this now as AI companies (and large companies of all kinds) rush to secure whatever advantages they can, regardless of the negative externalities. Meanwhile people who actually care about doing the right thing get trampled.
We need to shift the incentives by adding ruinous penalties for things that are currently quite commonplace if they are done by large players. Some dude training his own AI on his own computer can scrape and train. The fine for OpenAI or Meta using a single copyrighted book without permission should be in the tens or hundreds of millions.
What we're seeing currently in our society is a "loophole inversion" where the rules have an effect mainly via their loopholes. The most profitable activity is to find loopholes and exploit them as frenetically as possible to gain as much advantage as you can before the loophole is closed, or get people hooked on the loophole so it's retroactively legalized. Entities that are big enough to do this are big enough because they have lots of money behind them. Entities doing the same kinds of things without lots of money are not really doing much harm. So the best approach is to adopt a "sliding scale" in which even tiny violations by wealthy actors result in penalties enormously greater than fairly large violations by small players.
You aren’t being cynical, you’re arguing with disingenuous entities with money on the line. We already know it’s used primarily for the negative case. Everyone who ever intended High School or College knows this.
This is my thoughts as well! It makes smarter people smarter and dumb people dumber!
All at the cost of ... {List of negatives regarding the construction and powering of AI data centers }
Most importantly, it makes investors think dumb people are smart.
There is zero evidence for this.
LLMs make significant mistakes frequently and smart people have no way of judging those mistakes outside their domain expertise. They are also sycophantic and great at being an echo chamber which makes people feel smart even if they are not.
So I think the burden of proof is on you to prove that they somehow amplify intelligence, it seems highly unlikely.
Most domains have some kind of internal consistency/theory building you can do. A smart person can certainly notice inconsistencies when trying to learn something. In fact they're likely to be points of confusion that the smart person will dive into just to try to make sense of things, even if they don't suspect the LLM is at fault.
It helps smart people be barely more effective and helps dumb (more importantly, people who do not value effort, people who are lazy, people who are self absorbed) people shit out endless streams of worthless tokens that can swamp out everything.
Raising the noise floor like this only makes it that much harder to find "Smart" people, which we were already doing terrible at.
I use Claude every single day, but this is such a bad tradeoff. Maybe it will help me standup a quick fix when that is needed. Maybe it can help me dig through documentation to find relevant bits and figure out the unstated assumptions underlying it. Maybe it helps me generate test cases.
Meanwhile, my day to day life is now noise. All social media is noise. All content is noise. Slop pours onto me from all directions. Writing more test cases isn't helping me.
Am I smart? Am I dumb? I don't care, right now I'm deafened
All I've noticed is AI creating a perverse incentive to make everything as complicated and bureaucratic as possible, so only the people that know how to leverage AI to cut through it ever succeed.
I'm using Claude at work myself and am impressed with the product, but notice that this is the only reason I need to use it at all. Our product pages were shit to begin with, now they're AI-generated and somehow even worse. Our procedures are incomprehensible spaghetti with enough arbitrary context switching to give a sadistic Soviet municipal administrator an erection at the thought of watching anyone try to actually follow them.
Use AI to create inefficiencies, then use AI to bypass them. Those who can't do the latter will struggle to survive.
> Raising the noise floor like this only makes it that much harder to find "Smart" people, which we were already doing terrible at.
Clarification: to value “smart” people, which we were already doing terrible at.
Moneyshot: "The results are eye-opening. After six months, pupils using ai saw their average homework score rise by 18% across all subjects. The time they took to complete each assignment fell from an average of 64 minutes to 45. But come exam time, the same students scored 20% below their classmates who had not called on ai’s help. Homework scores once predicted exam performance; now those who score highest are, perversely, more likely to do worse in exams."
Study design: "David Stromberg of Stockholm University and Victor Lei and Wu Yanhui of the University of Hong Kong set out to fill the gap. They tracked 27,000 pupils aged 12-18 in China, where ai adoption has been fast. Around 80% reported using models such as Doubao and DeepSeek; the other 20% formed the control group."
> Homework scores once predicted exam performance
As with most training the journey is the point, not the destination.
That said, I think smart use of AI could help. It could explain concepts in a way that might help you understand better, it could probe your knowledge in a more dynamic way by tailoring questions, and so on. This requires the AI be restrained by some harness, not free to write down the answers for you.
> help. It could explain concepts in a way that might help you understand better
People say that all the time, but does anyone really think a lack of good explanations for things is a limiting factor in 2026? Or even 2010?
> This requires the AI be restrained by some harness
How would this magic harness look like and why would anyone use it?
Taking a forklift to the gym increases the number on the bar, but mysteriously your gains go down.
To extend this analogy:
Using the forklift as a spotter and to assist in loading weights increased gains. Then again, a human can do all those things, and provide real human connection.
I used to skip homework in high school because it was boring and cut into my personal life outside of school (including my job) but I would always do very well on tests.
Then I stopped caring about tests too.
I did the same thing through college. The only reason they passed me and I got a degree is that I built the school’s website and I built personal ecommerce sites for the head of art and his wife to sell their paintings.
I haven’t done shit since like 7th grade.
Worked at Facebook, Apple, Microsoft, same behavior there - did basically nothing for them while making thousands off my games on App Store.
Fuck authority
It is stopping some of the adults I work with from learning.
In the case of some management, is causing them to unlearn, forgetting about proper review and maintenance practices.
https://archive.is/w1eng
I know this study is focused in China but one thing I'd like to understand is the nuance in what kind of student is likely to always reach for ai for homework help as I think (especially in the US, can't speak for other countries) the varying powers that be that can determine the quality and type of education you can receive (does your country have means to buy textbooks per student or shared, do you have funds to give students ipads to take home or no, what sets the bar on how far curriculums can go, are teacher and faculty pay and incentives simply tied to student pass rates, etc.) and those do more in shaping what "use AI responsibly" will look like and why it's so different.
Use AI for homework, you will get it done faster and get a higher grade, and then get crushed on the exam.
i’m pretty sure there’s already a correlation between technology in the classroom and worse learning outcomes