Answers are grounded in this paper and cite their sources.
Generative Artificial Intelligence (GAI)–based intelligent tutoring systems represent a transformative innovation in school education, offering unprecedented learning personalization. This systematic literature review analyzes 53 scientific studies to identify the opportunities that GAI provides for the design and development of these tutoring systems. The findings show the predominance of general-purpose tools such as ChatGPT and models such as GPT-4, highlighting their ability to generate interactive content and adaptive feedback. Key features are identified, including adaptability, dialogic interaction, and emotional engagement. However, the analysis reveals critical gaps between current technical capabilities and deep pedagogical integration, including limitations in comprehensive student modeling, algorithmic opacity, and limited empirical evidence of long-term impact. In light of these gaps, strategic research and development opportunities are proposed, focusing on grounded instructional design, explainable architectures, operational mechanisms for ethics and equity, teacher co-orchestration tools, and rigorous efficacy studies. It is concluded that advancing the field requires a transdisciplinary approach to ensure that these intelligent tutors are pedagogically robust, ethical, and aligned with the goal of equitable and holistic education.
No takes yet. Share an insight, caveat, or question.
Plaza-Santillán et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: