The 2025 PISA scores are the lowest the OECD has ever recorded, and the alarm is justified. But “scores are down” is not an explanation. The report itself points at a specific set of culprits, and the most surprising one involves a tool students increasingly use to do their homework.
Every bad-news education headline eventually hits the same wall: it tells you the number fell without telling you why. The 2025 PISA release is more useful than that. Buried under the record-low averages is a fairly clear story about attention, reading, and technology, three threads that keep showing up together. Understanding them matters more than the ranking, because you cannot fix a decline you have not diagnosed.
The Short Version
The 2026 PISA results score declines are led by reading, which fell about 28 points across OECD countries since 2015, followed by math down 22, with science also slipping to record lows. The OECD links the slide to three connected factors: rising screen time and classroom device distraction, a sharp drop in reading for enjoyment, and heavy AI use. Notably, students who regularly used AI chatbots to draft essays or summarize texts scored around 20 points lower in science. The common thread is attention, the mental resource the falling skills depend on most. The causes point less to curriculum and more to how students spend their focus.
Table of Contents
Which skills fell the most
The declines are not uniform, and the pattern is a clue in itself. Reading fell hardest, down roughly 28 points across OECD countries between 2015 and 2025, which the OECD equates to about a year and a half of schooling. Math dropped about 22 points, and science also fell to a record low, per the OECD’s own report.
Why does the reading result matter most? Because reading is the skill that most depends on sustained, uninterrupted attention. A math problem can be chunked into steps. A long passage has to be held in your head all at once. When reading is the steepest decline, it points toward attention as the underlying problem, rather than gaps in any one subject’s teaching.
The screen-time connection
The clearest thread in the report is distraction, and it is measurable rather than anecdotal.
More than one in four students said classmates were distracted by digital devices in most or every science lesson, and those students performed worse and felt less connected to school, as coverage of the findings noted. This is not a claim that a phone in a pocket lowers a test score by itself. It is that a classroom where a quarter of students are half-watching a screen is a worse learning environment for everyone in it, including the students paying attention.
The reading-for-pleasure collapse
Alongside the screen data sits a quieter number that may matter just as much: reading for enjoyment has fallen sharply. This is the piece most likely to get lost, so let me back up and explain why it is a big deal. Decades of research tie voluntary reading to reading skill. Kids who read for pleasure get better at reading, almost regardless of instruction, because they simply do more of it. When that habit collapses, the skill it feeds collapses with it. The reading decline and the reading-for-pleasure decline are not two separate problems. They are the same problem measured two ways.
The AI-chatbot paradox
Here is the finding that surprised even people who follow this closely. Nearly half of students now use AI chatbots regularly, and the ones who used them to draft essays, summarize texts, and research topics scored about 20 points lower in science. These chatbots, marketed as a study aid, correlate with weaker results, at least the way many students actually use them.
The likely mechanism is straightforward once you say it out loud. Using AI to explain a concept can help. Using AI to produce the work skips the mental effort that builds the skill in the first place. If a student never wrestles with structuring an argument because a chatbot structures it for them, the muscle that PISA measures never gets exercised. The score gap suggests a lot of students are outsourcing exactly the thinking the test is checking for.
What the causes suggest
Diagnosis points to response. If attention, reading habits, and AI-use patterns are the drivers, the implications are fairly specific: protect focused time, rebuild voluntary reading, and teach students to use AI as a tutor rather than a ghostwriter. None of that is a curriculum overhaul. It is closer to changing how attention and effort get spent, which is harder to mandate but cheaper to try.
What To Know
- Reading fell hardest in PISA 2025, down about 28 points since 2015, with math down 22 and science at a record low.
- The report links declines to screen time, classroom device distraction, less reading for pleasure, and heavy AI use.
- Reading’s steep drop points to attention as the core problem, since reading depends on sustained focus.
- Students who used AI chatbots to do work scored about 20 points lower in science.
- The causes suggest protecting focus, rebuilding voluntary reading, and using AI to explain rather than to write.
Frequently Asked Questions
Which subject declined most in PISA 2025?
Reading fell the hardest, down about 28 points across OECD countries since 2015, which the OECD equates to roughly a year and a half of schooling. Math fell about 22 points and science also dropped to a record low.
Is screen time really causing lower test scores?
The report shows a strong association rather than proof of direct cause. More than a quarter of students reported classmates distracted by devices in most or every science lesson, and those students scored lower and felt less engaged. A distracted classroom hurts learning for everyone in it.
How does reading for pleasure affect PISA scores?
Voluntary reading is one of the strongest predictors of reading skill, because students who read for enjoyment simply read more. Reading for pleasure has fallen sharply, and that decline moves in step with the drop in measured reading ability.
Do AI chatbots hurt student performance?
The data suggests they can, depending on use. Students who regularly used chatbots to draft essays and summarize texts scored about 20 points lower in science. Using AI to produce work appears to skip the mental effort that builds the skill being tested.
What do the PISA causes suggest schools should do?
Focus on attention and habits rather than a curriculum overhaul: protect distraction-free learning time, deliberately rebuild reading for pleasure, and teach students to use AI to understand concepts instead of to complete assignments for them.
The Bottom Line
The 2025 PISA results are a warning, but the diagnosis matters more than the alarm. Reading fell hardest because reading needs the most sustained attention, and attention is exactly what screens, collapsing reading habits, and outsourced AI work all erode. The uncomfortable implication is that the fix lives less in the curriculum and more in how students spend their focus, at school and at home. For more on schools and technology, browse The Other Stream’s Education section, or our Tech coverage on AI and devices.
Related Posts
