Students are increasingly asking chatbots to do the one thing school is supposed to train them to do themselves: read, process, and argue from a text. That warning comes through clearly in a recent essay by literature teacher Tyler Jagt, who describes a student unable to get through a 20-page article that he says he once handled easily as an undergraduate. The problem is bigger than one distracted reader, and the evidence now stretching across test scores, classroom habits, and lab studies suggests higher education has a literacy problem that AI is happily making worse.
The numbers are ugly. A 2024 study cited in the piece found that 12th-grade students posted their weakest reading scores since tracking began in 1992, while almost a third landed below ”basic” performance. In fourth grade, 70% of students – about 2 million children in the United States – were said to read below the expected level. If that sounds like a future workforce problem, that is because it is one.
Why teachers say AI is making things worse
Jagt’s complaint is not that students use tools, but that they use them as a substitute for reading itself. Instead of wrestling with a long article, many ask AI to summarize it; instead of drafting an essay, they let a chatbot produce the first pass. That may look efficient on a deadline, but it turns coursework into outsourcing, and universities are hardly helping when they hand out access to advanced AI models through corporate partnerships.
This is where the optimistic AI pitch starts to wobble. One of the few large studies claiming ChatGPT improved learning was withdrawn in May, and the research cited by Jagt points in a darker direction: reduced critical thinking, weaker memory, and lower brain activity in students who use AI for creative assignments compared with those who rely on Google or no search tool at all. In one striking figure, 83% of students could not quote even a line from their own work.
The brain does not outsource for free
There is also a useful reminder from an older 2017 study: even a smartphone sitting nearby, face down, can dent cognitive performance. The broader point is obvious enough to irritate tech evangelists – attention is a muscle, and muscles weaken when they are never used. That is why the ”AI will free students for higher-level work” line sounds neat but often collapses in practice; if a student cannot follow a 20-page argument, there is no higher level to jump to.
Jagt argues that the reading crisis is structural, yet schools often treat it like a private inconvenience for instructors to patch up with extra explanations, whiteboard sketches, and split assignments. That is a bad bargain. The more coursework gets trimmed to fit AI-assisted habits, the less practice students get at the exact skills that school is meant to strengthen: sustained attention, recall, and actual comprehension.
What happens if classrooms keep adapting to chatbots
The uncomfortable question is whether universities will keep normalizing this behavior because it is easier than confronting it. If they do, the winners will be the platforms that save students time and the institutions that can pretend productivity has improved. The losers will be everyone who still expects graduates to read carefully, think independently, and write something that sounds like it came from a human brain.