AI and Educational Achievement: A Comparative Analysis of Turkey and Estonia 
 
This study examines whether AI-compatible digital infrastructure moderates the relationship between socioeconomic background and academic achievement. While artificial intelligence and digital technologies are often promoted as tools for reducing educational inequality, their distributional effects remain uncertain. Focusing on inequality as the strength of the socioeconomic gradient in achievement, the analysis investigates whether digital infrastructure reshapes this relationship. 

Using student-level microdata from PISA 2022, the study adopts a comparative framework involving Turkey and Estonia. Socioeconomic status is measured using the ESCS index, academic performance is captured through mathematics plausible values, and AI-related digital capacity is proxied by standardized indicators of school-level ICT infrastructure. Survey-weighted linear regression models with interaction terms are employed. 

The results show that socioeconomic status strongly predicts mathematics achievement in both countries. However, higher levels of digital infrastructure do not reduce inequality; instead, they are associated with a steeper socioeconomic gradient. These findings suggest that AI-compatible digital infrastructure does not automatically function as an equalizing force in education and that its effects are shaped by existing socioeconomic and institutional contexts.