Studies on How Generative AI Can Adaptively Respond to the Affective Needs of Learners: A Systematic Literature Review

Abstract

Generative Artificial Intelligence (GenAI) has become increasingly prominent in education, providing adaptive and personalized support for learners. While prior studies have emphasized cognitive benefits such as knowledge retention and problem-solving, the affective dimensions of learning — motivation, engagement, confidence, and emotional regulation — remain comparatively underexamined. This paper presents an updated systematic literature review (SLR) on how GenAI adaptively responds to the affective needs of learners. Following PRISMA 2020 guidelines, ten peer-reviewed core studies published between 2020 and 2024 were analyzed in depth, and their findings were triangulated against thirty additional supporting sources spanning affective computing, motivational theory, mental-health-oriented chatbot research, and AI governance, for a total of forty cited references. Findings reveal five major themes: affective signal detection and adaptive feedback, reinforcement-based motivation pathways, integration of GenAI with social-emotional learning (SEL) frameworks, personalization and scaffolding strategies, and system-level governance considerations. The review identifies persistent gaps, including a scarcity of causal evidence linking affective adaptation to long-term outcomes, heavy reliance on unimodal (mostly textual) affect-sensing techniques, and underexplored teacher readiness and ethical governance structures. The paper concludes by proposing a conceptual framework for affect-adaptive GenAI and recommending longitudinal, mixed-methods research to validate and extend its role in building emotionally intelligent digital learning environments.