Family-school autonomy support for children's responsible use of generative artificial intelligence: a self-determination theory synthesis and developmental research agenda.

Generative artificial intelligence is becoming embedded in children's and adolescents' learning, yet guidance often oscillates between uncritical adoption and restrictive control. This mini review applies self-determination theory to ask how autonomy-supportive guidance from families and schools might foster responsible use, and it audits how far the available evidence can carry that account. Of the 46 primary-data studies reviewed, 35 sampled university students and 8 included participants under 18; 26 were cross-sectional, and none measured family and school inputs in the same participants. Within those limits, current evidence suggests that educational outcomes depend less on frequency of use than on whether learners retain cognitive agency, engage reflectively, and regulate when and why they delegate tasks to artificial intelligence. Autonomy support may strengthen learning engagement by satisfying needs for autonomy, competence, and relatedness and by supporting intrinsic motivation, hope, self-efficacy, and self-regulated learning. It may also protect academic integrity by facilitating moral internalization, authentic self-attribution, and reasoned decision-making within clear institutional boundaries. Conversely, controlling or weakly structured environments may promote dependent cognitive offloading, unreflective use, and overreliance. Because direct evidence in minors is scarce, we treat the developmental claim as an agenda rather than a finding: we present family-side and school-side evidence separately, state their synergy as three competing and falsifiable hypotheses, and grade each proposed link by the strength of evidence behind it. We identify priorities for longitudinal, behavioral, cross-cultural, and participatory research.