Paradox of Mitigating the Achievement Gap: Digital Technology and Socioeconomic Inequality in PISA 2025

Digital technology has expanded access to educational resources. In this paper, we investigated the moderating effect of 12 ICT indicators across 5 categories (access, frequency, use, interest, AI usage) on the relationship between students’ socioeconomic status (SES) and academic achievement, using a cleaned PISA 2025 dataset of 335,424 observations across 44 countries. The baseline, additive, and interaction models were estimated using weights and controlling for country and gender. We found that a one-unit increase in SES is associated with a 30- to 32-point increase in students’ science scores. Varying moderating effects were shown from different ICT indicators: 4 negative interactions, 3 positive interactions, and 5 nonsignificant interactions. It further highlighted the importance of analyzing the multidimensional nature of ICT. A key finding of this research is that all 4 indicators with negative interactions show a negative association with science scores controlling for SES, country, and gender. This means the achievement gap is narrowing across different SES groups as ICT increases, but in a way that an ICT increase for high-SES students is associated with a larger science score decrease compared to low-SES students. However, interaction terms added relatively little explanatory power, and this study cannot conclude causation because of the survey nature of PISA. This paper also provides initial insight into large-scale data on AI use in education. We have found AI use for school work negatively associates with science scores and has a negative interaction term.