> IQ scores in most of the Western world began reversing sometime from the 90s to early 00s. [1] That link is for the US but similar stuff has also been observed in most of every Western/OECD nation.

I wonder to what extent demographic shifts account for the observed reversal? It would be interesting to tease out what happens where composition remains constant, i.e. Poland. Though I doubt this explains the entire effect.

> By contrast they're still rapidly increasing in the 'Global South'.

seems like the Flynn effect doing its thing just as it did in western countries in the 20th century.

I get the feeling you want to blame "woke" ideologies for this decline. Please correct me if I'm wrong.

In my opinion we just have it too easy nowadays. Too much entertainment, screens and not enough challenges to keep kids thinking they need to be as sharp as they can be.

Those top scorer countries just soooooo happen to have world lowest birthrate. Maybe it has to do with extraordinary pressure to succeed at all cost since one is baby?

I see news and forums claiming the problem is screens/smartphones and social media.

Given the big gap between the east and the west... does anybody know if these are somehow less used/abused in asian countries?

Any other reason why this gap might happen?

Thanks

As they say, demographics is destiny.

Sure you can torture your children with ever harsher "education", almost none of which matters in non-research positions. Then don't be surprised when they run out.

> So much for all the EdTech startups that want you to use their AI teachers.

The AI they’re talking about is not EdTech AI-based teaching tools. It’s students who use ChatGPT to handle work for them so they can ignore any teaching parts.

Apples and oranges. Using this study to draw conclusions about AI EdTech would be a mistake

Sloppy conclusion about edtech startups.

If everyone is hiring a "tutor" to just do their homework for them, that doesn't mean that all tutors are bad.

> So much for all the EdTech startups that want you to use their AI teachers.

Kid pasting the question into ChatGPT and pasting its answer back is not the same as an AI teacher persona.

In my mind an AI teacher can be like a traditional education package (syllabus, material, questions, keeping track of progression) but way more flexible. I don't see no reason why AI teachers shouldn't give Bloom's 2 sigma effect to kids (and adults) very soon.

The "AI use ... To help me learn" graph is really interesting! It may be interpreted as showing moderate selective AI use being beneficial and, perhaps, even mitigating the effect of socioeconomic status, that is democratising the learning of science. However, it may also indicated that groups with low AI use/access are left behind, or at least that over-reliance on AI in learning is detrimental.

The breakdown for AI usage is four categories though, not three. While the first three are negatively correlated with increased usage, the one for using AI "to help me learn" is much more flat when it comes to usage frequency.

What percentage of the population is expected to get university education today vs the sixties?

What's the difference in how many people apply?

> The exam has not changed in difficulty

Has what students are expected to know changed since the 60s? For instance, have there been any, I dunno, new developments in technology over the past 60-odd years?

>- Social media and doomscrolling without sufficient parental controls: Several OECD countries are already addressing this by introducing or considering restrictions on teenagers’ access to social media.

Surely this time banning things will improve test scores!

Even before you get to the scores in the report, it shows what percentage of students come from schools that have banned phones. There doesn't seem to be any obvious correlation where banning phones from school gives better results. Eg Estonia averaged 520 points with 26% of students from schools that ban phones. 88.9% of students were in NZ schools where cellphones are banned, but NZ averaged 511 points.

I think before we start banning more "random" things we should consider whether these bans actually produce results. Let the countries that are making these changes run their banning experiment and the rest of us should wait and see whether their scores meaningfully improve afterwards.

Take over Scotland and Wales?

Interesting. On page 34 of the report there's this:

> Hasty responses (percentage of responses that are both incorrect and fast – average across reading items)

Spain is at 9.7%, which is a bit over 8.9% oecd average.

> in Spain kids don't take these tests seriously

Is there evidence for this attitude not generalizing?

Last I heard it started in the 90s. Although a lot of horrifying graphs do have their inflection point the year the iPhone came out.

If you're referring to the few-point difference between 2012 and 2018 in the first chart here: https://www.oecd.org/en/publications/pisa-2025-results-volum... I think that's more likely to be statistical noise. Keep in mind that PISA was originally normed to a mean of 500 and standard deviation 100. Even the more obvious drop after 2018 (I blame the pandemic) down to a mean of 466 on the reading test in 2025 isn't all that huge in absolute terms, though definitely statistically significant with the large sample they have.

A factor, certainly, but culture, including the culture around education, are arguably bigger factors (which social media is also a vector for modifying; medium is the message and all that).

Social and cultural expectations and conventions are extremely important. Even in supposedly laissez-faire Western countries where traditional cultural expectations have withered to the point of laxity, people still live their lives according to convention and expectation, even if those conventions and expectations are increasingly vulgar and dictated by powerful interest. (Indeed, the fertility crisis is determined largely by patterns of convention. It used to be that, for most people, it was expected that you would marry and have children earlier rather than later. Now, our cultural expectations actually penalize that kind of behavior. You're supposed to "find yourself", "play the field", or "build up your career" first.)

Human beings operate more by convention than by moral reasoning. Society needs an environment demands good things from them. Even when we know what the right or good course of action is through moral reasoning and discernment, it is very helpful for social convention and culture to support those choices.

The law is also traditionally a source of guardrails that help people attain better lives.

> The U.S. scores are interesting.

Unfortunately, those scores are not reliable. It is a recurring problem with the USA for the PISA studies.

The preliminary chapter explains why data has lower confidence is some places. For the USA, there were many problems, including political reluctance in some states or schools, that lead to uncontrolled biases. Notably, the ratio of unreachable children is 40 %, twice more than their upper limit for valid results.

Parents used to offload parenting to streets, fields and forests for decades. I don't think the offloading is the cause. But switching the offloading provider to SOCIAL MEDIA (of which Youtube is one) might've been.

The modern trend is parents spending a lot more time and involvement with their kids then used to be the norm. The norm was that kid were outside unsupervised doing whatever. The expectation on parents went up rather then down.

> is doing math at the level of the best mathematics

Terence Tao a few days ago [1]:

> When we point the AIs at really difficult problems where none of the standard techniques apply, they are still very, very bad, I mean they're just randomly guessing.

Please don't reply saying they'll be proving the Collatz conjecture next week. It adds nothing.

[1] https://youtu.be/svl_1upFpQo (worth watching on the role of LLMs in maths)

The decline started in early to mid 2010's, this has little to do with AI.

> Makes sense. The kids can see that education isn't really going to be a sellable skill when AI ends up coming into full force.

If true, this is a disastrously foolish thing to believe. Education is only going to become more important as supervising increasingly advanced AI agents becomes more difficult. We need people who understand all different domains of human endeavor at the deepest level possible.

I mean, what's the alternative? AI does everything for me and I play Fortnite all day? Might as well plug yourself into the Matrix.

Dunno man, the articles ai writes are empty cardboard crap, and the code it produces is hot garbage. “Cleaner upper of AI messes” might be a good future skill to have.

I hope people recognize this account doom posts what amounts to misanthropic propaganda at a very high rate. There’s never any question how the comment is going to be framed. It’s always, “give up little humans, your robot overlords are here.”

Yes, kids are the demographics known for long term thinking and rational decision making.

Where do you think they are putting their energy?

That's quite funny, but the new recruits at my work place are super smart and full of piss and vinegar. Much more motivated than some of the older workers. The problem is that they will work lots of unpaid overtime, and they have no idea about work boundaries and being exploited by employers.

on the other hand, they'll be taking care of us when we're old :(