Every time students' test scores fall, people blame whatever's newest — phones, COVID, now AI — but is that the real cause?

This explores why each new technology becomes the leading suspect when international test scores fall. First, a limit: the collection holds no PISA studies, no international testing data and nothing that compares the phone, COVID and AI explanations. It can't say which of them, if any, is right. What it does have is material on how explanations like these get assembled, and that turns out to be the more interesting part of the question.

The closest match is Romero's argument that AI getting better at elite mathematics and humans getting worse at basic numeracy are one curve measured at opposite ends, not two separate stories Are AI's math gains and human math losses really connected?. The note is open that this pairing is rhetorical and provocative, not causally proven. That is the pattern behind rotating blame. A decline is observed, a big new technology arrives at about the same time, and putting the two side by side makes a story that feels complete without anyone showing that one caused the other. When the next technology shows up, it slots into the same story.

A less obvious connection comes from research-integrity work. One demonstration had language models write 288 complete finance papers from 96 statistically significant patterns, each with its own invented theory explaining the result after the fact Can AI generate hundreds of fake academic papers automatically?. Researchers call this HARKing: hypothesizing after the results are known. Public arguments about PISA run on the same logic. The result (scores fell) is fixed first, and a plausible cause is found afterward. Since a convincing reason can be written for almost any pattern, the fact that a story is convincing tells you very little. This may explain why the explanations rotate instead of building on each other. None was tested hard enough to be ruled out, so each simply gets replaced.

That doesn't mean the worry is empty. A four-month EEG study found that brain connectivity and memory of one's own work dropped as people relied more on an AI assistant Does AI assistance weaken our brain's ability to think independently?. But that is 54 people doing an essay task in a lab, not a national trend among teenagers. The gap between a controlled study and a country's falling scores is exactly where rotating stories grow. Another study warns against a related move: treating group differences as noise to explain away. It found that Indian writers accepted more AI suggestions than American writers, and its authors argue this reflects culture, not a bias to correct Is higher AI use by Indian writers a confound to control?. By the same reasoning, countries' PISA paths may differ for local reasons that a single global culprit hides.

Finally, there is a reason the AI version of this story may spread especially easily. We have learned to discount advertising because we know it comes from an interested party, but there is no settled habit yet for reading AI-generated or AI-focused discourse How do we learn to read AI-generated text critically?. The phone story got heavily criticized over time, while the AI story still arrives without that kind of built-up skepticism. So the corpus can't tell you what caused the PISA decline. What it can tell you is why the answer keeps changing: decline-and-technology stories are cheap to write, hard to disprove, and judged by how much they resonate, not by evidence.