Declining Academic Performance Among Students Linked to AI Addiction

The academic abilities of adolescents in various countries have come under scrutiny following the latest Programme for International Student Assessment (PISA) results, which show a decline in student performance across several key subjects.

According to reports, the PISA 2025 results indicate the lowest average performance ever recorded in OECD countries for science, reading, and mathematics. This assessment involved over 760,000 15-year-old students from 91 countries and economies, representing approximately 33 million adolescents worldwide. PISA remains one of the largest international surveys used to measure the ability of 15-year-olds to apply their knowledge and skills in reading, mathematics, and science.

The sharpest decline was observed in reading abilities. OECD data shows a significant drop over the last decade; in OECD countries, average reading scores fell by 28 points between 2015 and 2025, while mathematics scores dropped by 22 points. The OECD estimates this decline is equivalent to approximately one and a half years of learning for reading and more than one year for mathematics.

Meanwhile, science proficiency remained relatively more stable between 2022 and 2025 compared to reading and mathematics. However, overall, PISA 2025 still recorded the lowest average OECD performance across all three fields. Approximately 20 per cent of 15-year-old students in OECD countries are now categorised as low-performing in all three subjects, an increase from 16 per cent in 2022.

A particular concern is the shift in reading proficiency. The OECD found a decline in skills that are increasingly vital amidst the heavy flow of digital information, such as evaluating information, connecting various sources, and exercising critical thinking regarding what is read.

The use of Artificial Intelligence (AI) by students has become a focal point of analysis in the PISA 2025 study. The OECD emphasised that the results do not suggest that AI is solely responsible for the decline in student achievement. The relationship between AI usage and learning outcomes is considered complex and highly dependent on how the technology is utilised.

PISA data shows that students who use AI for specific school tasks, such as summarising texts, conducting initial research, or drafting essays, achieved, on average, lower science scores compared to students who did not use it for those purposes. For certain activities, the difference reached approximately 20 points.

However, these findings cannot be directly interpreted to mean that AI causes low grades. The OECD highlighted various factors influencing this relationship, including how students use AI and the characteristics of the students adopting the technology. Conversely, purposeful use of AI can be linked to better learning outcomes. Students who use AI regularly for general purposes and have opportunities in school to learn AI literacy tend to have slightly higher science scores than those who do not receive such instruction.

The findings suggest that the issue of AI use in schools is becoming more complex. The question is not merely whether students use AI, but for what purpose and how the technology is employed. The OECD noted that limited or moderate use of digital devices for learning activities is often associated with better results. In contrast, excessive use and digital distractions in schools are linked to lower outcomes.

Approximately 28 per cent of students in OECD countries reported that their peers are distracted by digital devices during most or all science lessons. Students in classroom environments with high levels of digital distraction also tend to show lower levels of learning engagement. Consequently, the OECD emphasised the importance of ensuring that technology and AI are used as tools to support learning, rather than as replacements for a student’s own thinking process.