THE ROLE OF GENERATIVE ARTIFICIAL INTELLIGENCE IN FOSTERING SELF-REGULATED LEARNING STRATEGIES WITHIN DIGITAL HIGHER EDUCATION

The rapid spread of generative artificial intelligence (GenAI) tools such as ChatGPT has created both opportunities and risks for self-regulated learning (SRL), a competence that largely determines student success in digital higher education. This article aims to synthesise current evidence on how GenAI influences students' SRL strategies and to identify the pedagogical conditions under which GenAI fosters rather than substitutes self-regulation. An integrative literature review was conducted on 18 peer-reviewed publications from 2023–2026, comprising nine empirical studies involving 1,394 university students, six systematic reviews and meta-analyses, and three conceptual design frameworks. The publications were coded deductively against Zimmerman's threephase SRL model and analysed thematically. The results show that GenAI supports all three phases: goal setting and resource search in the forethought phase, problem-solving strategies and progress monitoring in the performance phase, and feedback and self-assessment in the self-reflection phase. However, gains in task performance did not consistently translate into knowledge gain, transfer or accurate self-evaluation, and several studies reported reduced mental effort, over-reliance and metacognitive laziness. Improvements in SRL processes were observed when GenAI use was combined with explicit metacognitive support or with reflective feedback based on learning analytics. The article proposes a conditional model in which GenAI acts as a regulation amplifier only when the learner retains metacognitive agency, and as a regulation substitute otherwise. Practical recommendations are formulated for universities, including those of Uzbekistan, concerning assignment design, AI literacy and teacher development. The findings can inform institutional policies on the responsible integration of GenAI into digital higher education.