AI Use Surges as Teenagers Lose the Habit of Deep Reading

The share of “hasty readers”—students who rush through written texts and provide immediate but incorrect answers—nearly doubled between 2018 and 2025, according to data from the Organization for Economic Cooperation and Development (OECD). The decline in deep reading engagement is occurring as 86% of 15-year-old students across participating OECD countries report using AI chatbots for at least one school-related task.

The findings come from the OECD’s Program for International Student Assessment (PISA), which measures how well 15-year-olds near the end of compulsory education can apply knowledge to real-world challenges. Established by the Paris-based intergovernmental organization in 1997, PISA has been administered every three years since 2000 across mathematics, reading, and science.

Approximately 20% of students across OECD member nations—one in five—say they use AI chatbots almost every day for schoolwork. After socioeconomic status is controlled for, these daily users consistently scored lower in science than peers who used AI less frequently or not at all. The 14% who reported never or almost never using AI chatbots for schoolwork maintained higher baseline academic performance than the general population of AI users.

The assessment also found higher average science scores among students using AI chatbots on a limited schedule, from once a month to once or twice a week, than among both non-users and daily users. This “Goldilocks effect” was most visible when AI served as a supplementary tool for preliminary tasks, including summarizing assigned reading passages or conducting initial background research on unfamiliar topics. When used weekly for those targeted functions, performance matched or slightly exceeded that of non-users after socioeconomic variables were adjusted.

Researchers warned that heavy reliance on AI can create a damaging feedback loop for developing minds. Students who routinely allow automated systems to handle reading comprehension, text summarization, and logical synthesis miss cognitive opportunities to build independent reasoning skills. Students with weaker critical reading abilities then struggle to detect incomplete, biased, or hallucinated information generated by AI platforms.

“Most troubling is the decline in the very skills that matter most in the AI age: evaluating information, making connections across multiple sources, and thinking critically about what we read,” the OECD report stated. “In a world increasingly flooded with information, our students are less likely to think deeply and separate signal from noise.”

The OECD said the metric reflects how students interact with information rather than a direct negative impact from the technology itself. Comparing mental development to physical exercise, the report noted that learning requires active engagement rather than passive consumption: “In the same way that we do not become fit by watching sports but by doing sports, learning does not occur through the consumption of content,” the report concluded, “but as a productive cognitive struggle of the mind with new material.”

Recent PISA cycles have tracked broader structural shifts in student learning. The PISA 2022 study involved nearly 700,000 students across 81 participating countries and economies. The OECD emphasized that declines in reading, mathematics, and science began before consumer AI tools emerged; standardized PISA scores across member states have followed a long-term downward trajectory dating back to 2003, driven by factors including changing media consumption, reduced instructional time, and learning disruptions linked to the COVID-19 pandemic.

generative AI entered education debates rapidly after the public launch of OpenAI’s ChatGPT in late November 2022. Its ability to draft essays, solve complex equations, and generate computer code prompted immediate policy debates across school systems globally.

Several major U.S. school districts initially blocked access to AI platforms on school networks and devices in early 2023 because of concerns about plagiarism and academic integrity. They included New York City Public Schools, the nation’s largest public school system, as well as districts in Los Angeles and Seattle. Most districts rolled back outright bans by late 2023 and 2024, moving toward instructional integration guidelines.

Federal and international guidance also shifted toward structured implementation. In May 2023, the U.S. Department of Education’s Office of Educational Technology released a policy report advocating a “human-in-the-loop” model, asserting that AI tools should assist teachers and students without replacing human interaction or cognitive effort. UNESCO followed in September 2023 with global guidelines for generative AI in education, recommending an age floor of 13 for independent chatbot usage and urging school systems to safeguard foundational literacy against automated shortcuts.