Self-Directed Generative AI Use and Academic Development Among Secondary School Students: A Large-Scale, One-Year Longitudinal Study

Many students use generative AI (GenAI) tools for schoolwork on their own, without pedagogical guidance. Yet it remains unclear whether such use supports or undermines their academic development, or which students turn to it. We examined both questions using TOSCAneo data from 4,062 students in 136 secondary schools in Germany. Data were collected at the beginning and end of the 2023/2024 school year and included standardized achievement tests and measures of academic self-concept, value beliefs, and effort across six subjects. Students’ reported use of GenAI tools for academic activities formed two related factors: task-completion use and learning-oriented use. GenAI use increased over the school year for both task-completion use (d = 0.35) and learning-oriented use (d = 0.33), with only 5% of students never using GenAI. In cross-lagged panel models, lower prior achievement predicted more subsequent GenAI use in all six subjects for task-completion use and in five of six for learning-oriented use. Neither form of use predicted subsequent achievement. A use profile characterized by more task-completion and less learning-oriented use predicted lower subsequent effort in four of six subjects. No comparable pattern emerged for self-concept or value beliefs, and prior self-concept, value beliefs, and effort did not consistently predict later GenAI use. Overall, rather than predicting later achievement, self-directed GenAI use was related to which students turn to these tools and how much effort they invest. Guidance may therefore be particularly important for lower-achieving students, who use GenAI more than their higher-achieving peers.