Answers are grounded in this paper and cite their sources.
Agentic AI systems are self-contained, goal-based entities designed to operate autonomously with little human intervention and dynamically respond to shifting contexts. This survey considers their transformative potential in education, ranging from intelligent tutoring systems and adaptive test-taking platforms to autonomous learning companions that can provide customized guidance and assistance. We provide an organized overview of the state of the field, classify existing and developing uses and assess their pedagogical worth, benefits and trade-offs. In the process, we also refer to such areas as learner motivation, equity, transparency and ethical concerns. Finally, we chart directions for future research and ethical implementation, highlighting how autonomous AI can reshape teaching, learning and educational spaces within the next several years.
No takes yet. Share an insight, caveat, or question.
Kostopoulos et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: