Teenagers' reading skills fall to record low - study

The reading skills of 15-year-old students in OECD countries have deteriorated sharply over the past decade and have dropped to the lowest level in the history of the PISA study. The organization connects the negative trend with a complex of factors, including the growth of screen time, digital distractions, and the peculiarities of using artificial intelligence in education.

This is evidenced by the results of PISA 2025, published by the Organization for Economic Co-operation and Development on September 8.

This was the largest PISA study in history: over 760 thousand 15-year-old students from 91 countries and economies participated in the testing. The sample represents approximately 33 million teenagers worldwide.

On average across OECD countries, reading results between 2015 and 2025 fell by 28 points, which the organization equates to about one and a half years of schooling. Mathematics indicators for the same period decreased by 22 points - just over one school year. Among countries and economies that are not members of the OECD, the drop in reading was 16 points.

Only between the last two PISA cycles did the average reading score in the OECD decrease by another 14 points - this is the largest drop between two rounds of the study in its entire history. Currently, one in five 15-year-old students in OECD countries demonstrates low results simultaneously in reading, mathematics, and science. In 2022, this was about one in six.

OECD paid attention to the use of AI

A separate block of PISA 2025 is dedicated to digital technologies and artificial intelligence. OECD emphasizes that the relationship between AI use and attainment is complex and depends on which specific tasks and in what manner students use it.

On average, 46% of students in OECD countries reported that they use AI chatbots for learning at least once a week. At the same time, schoolchildren who used AI for writing texts, summarizing materials, or conducting research on average showed lower results than those who did not use such tools for these tasks.

In particular, students who stated that they do not use AI for preparing texts of written works showed better results than those who used it for writing drafts. Meanwhile, among active AI users, higher results were demonstrated by schoolchildren who were taught to assess the reliability and quality of information generated by artificial intelligence.

Thus, the OECD does not conclude that the spread of AI itself caused the overall decline in results. The organization believes that technologies can both help learning and worsen its results depending on the manner of use.

Students spend more time in front of screens

Another problem cited by the OECD is digital distractions. More than a quarter of students reported that their classmates get distracted by digital devices during most or every science lesson.

Students in such classes on average showed lower achievement, were less engaged in the learning process, and felt a weaker connection with the school.

According to OECD data, since 2022 the share of teenagers who spend more than an hour a day on social networks has increased by nearly six percentage points. In parallel, the organization notes a growth of so-called hasty reading, when a student quickly skims a text but misunderstands it. The share of such cases has almost doubled compared to 2018.

AI is used more actively in Ukraine than the OECD average

PISA 2025 included students from 17 out of 27 regions of Ukraine. Of them, 54% reported using AI chatbots for learning at least once a week, while the OECD average is 46%.

For summarizing read texts, AI was used by 42% of surveyed Ukrainian students against 30% on average in the OECD, and for writing texts - 43% against 29%. For preliminary search of information on a new topic, AI was used by 37% of Ukrainian teenagers with the OECD average of 31%.

The decline in educational outcomes has economic consequences

OECD emphasizes that the problem goes far beyond school education. Weaker basic skills mean lower future labor productivity, fewer opportunities for social mobility, and more difficult adaptation of workers to technological changes.

OECD Secretary-General Mathias Cormann noted that technology and AI can enhance learning and prepare young people for the future, but only when used purposefully, and not as a substitute for attention, effort, and understanding of the material.

Sources: OECD, PISA 2025, The Wall Street Journal