OECD: AI Use Tied to Lower Student Scores — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OECD released findings from its 2025 PISA assessment of over 760,000 students across 91 countries, showing that general AI use correlates with lower academic performance in science after adjusting for socioeconomic status.
- Students who used AI daily or only once or twice a year performed worse than non-users, while weekly or monthly users saw less of a decline, indicating frequency plays a role in academic outcomes.
- Students who used AI for task-specific purposes like summarizing texts or drafting assignments scored relatively lower, whereas those using it for preliminary research or general learning support saw smaller performance gaps.
- Students who were frequently asked to assess the quality of AI-generated information outperformed non-users, especially when using AI weekly to help them learn, suggesting critical engagement enhances benefit.
- AI use is more prevalent among students from advantaged backgrounds, likely due to better access to devices, internet, and paid tools, and varies widely by country—95% of Vietnamese students reported use versus 60% in Japan.
- Andreas Schleicher, OECD director of learning and skills, emphasized that learning advances when technology enhances cognitive struggle, not when it replaces the mental effort required to engage with new material.
Why it matters: Weekly AI users trained to evaluate AI output outperform non-users, meaning schools that integrate critical assessment into AI use could turn a liability into an advantage. The 35-point gap in usage between Vietnam and Japan highlights unequal access, risking deeper educational divides where AI adoption favors wealthier students.
Ask SkimNews
