AI-Assisted Adaptive Learning in Education: Evidence, Teacher Competency, and Implementation Conditions

Abstract

Artificial intelligence (AI) is increasingly being integrated into educational technologies that personalize practice, generate feedback, and support teachers’ instructional decisions. This paper examines the conditions under which AI-assisted adaptive learning is most likely to improve educational outcomes without weakening teacher agency or learner autonomy. The study uses a structured evidence-synthesis design based on recent international guidance, a 2026 World Bank systematic review and meta-analysis of randomized controlled trials, and the OECD TALIS 2024 country evidence for Uzbekistan. The synthesis indicates that adaptive and AI-enabled interventions are associated with a positive average learning effect of 0.125 standard deviations, while AI-powered tutoring or instruction shows an average effect of 0.12 standard deviations; however, the evidence base remains limited and does not establish that generative AI is superior to earlier adaptive technologies. Uzbekistan provides a particularly relevant implementation case: 62% of surveyed teachers reported using AI in their work, while substantial shares identified infrastructure and skills as constraints. The paper proposes an implementation model that combines evidence-based tool selection, teacher AI competency, formative assessment, human oversight, and continuous evaluation. The main conclusion is that AI should be deployed as a pedagogical augmentation layer rather than as a substitute for teachers, with effectiveness judged by learning outcomes and equity rather than by technology adoption alone.

Keywords

adaptive learning; artificial intelligence; teacher competency; learning outcomes; educational technology

Authors & affiliations