OECD Report: Students Using AI Generally Score Worse, But Critical Assessment Improves Outcomes

AI Use and Academic Performance

A recent report from the OECD's Programme for International Student Assessment (PISA) reveals that students who incorporate AI into their studies tend to achieve lower academic scores than those who do not. This finding is based on data collected in 2025 from over 760,000 15-year-old students in 91 countries, marking the first PISA study since AI became widely accessible.

Varying Impacts of AI Application

The study found that the impact of AI use varies depending on its application. Students who used AI for specific tasks like summarizing texts or drafting assignments performed relatively worse. Conversely, those who used AI for preliminary research or general learning support experienced a smaller performance drop. Frequency of use also played a role, with both very infrequent (once or twice a year) and daily AI users performing the worst, while monthly or weekly users showed better results.

The Role of Critical Assessment

The report highlights a crucial nuance: AI use combined with training in critically assessing AI-generated information can lead to improved outcomes. Students who regularly used AI for general learning and were frequently asked to evaluate the quality of AI output outperformed non-users. This suggests that 'moderate and intentional use of AI for schoolwork and learning' can have positive effects when paired with critical thinking skills.

Cognitive Struggle in Learning

Andreas Schleicher, OECD director of learning and skills, emphasized that learning occurs through 'productive cognitive struggle' rather than passive consumption of content. He noted that technology can enhance this struggle, leading to student advancement, but can also 'short-circuit' it, undermining learning if not used thoughtfully.