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 a global educational report based on 2025 PISA data from 760,000 students across 91 countries, showing a negative correlation between AI use and academic performance in science.
- Students who used AI for task-specific purposes like summarizing texts or drafting assignments scored worse than nonusers, even after adjusting for socioeconomic status.
- Weekly AI users who applied the technology broadly to 'help me learn' outperformed nonusers slightly, especially when taught to assess AI-generated content critically.
- Frequent AI users — both daily and those using it only once or twice a year — performed worst academically, while monthly or weekly users fared better, indicating frequency and intent matter.
- AI use is more prevalent among students from advantaged backgrounds, reflecting disparities in access to devices, internet, and paid tools, with over 95% of Vietnamese students using AI versus 60% in Japan.
- Curiosity levels are highest among daily AI users, but this engagement does not currently translate into improved academic performance.
Why it matters: The data reveals that unguided AI use widens achievement gaps, as students with fewer resources gain less benefit despite lower access. Meanwhile, structured training in evaluating AI output allows regular users to outperform peers, suggesting equity-focused digital literacy programs could shift outcomes materially.
Ask SkimNews
