Toward Human-Centered Agentic AI in Education: A Systematic Review and Multi-Layer Integration Framework
DOI:
https://doi.org/10.61159/mgz7jn37Keywords:
agentic AI, artificial intelligence in education, systematic literature review, personalized learning, adaptive learning, machine learning, natural language processing, educational technology, PRISMA.Abstract
The rapid integration of agentic artificial intelligence (AI) in educational environments has generated growing interest in its potential to transform teaching and learning processes. This systematic literature review (SLR) synthesizes findings from 75 peer-reviewed studies published between 2025 and 2026, identified through Scopus, Web of Science, and IEEE Xplore databases, following PRISMA guidelines. The review examines the applications, dominant technologies, performance outcomes, and ethical challenges associated with agentic AI in education. Findings reveal that agentic AI systems characterized by autonomy, goal-directed behavior, and adaptability are increasingly deployed in personalized learning, adaptive assessment, and collaborative learning environments. Machine learning and natural language processing emerged as dominant enabling technologies. Despite notable improvements in student engagement and learning outcomes, significant challenges persist, including data privacy concerns, implementation costs, insufficient educator training, and ethical risks such as algorithmic bias. A novel multi-layer conceptual framework is proposed to guide the ethical and pedagogically grounded integration of agentic AI. Future research directions are outlined across short-term, medium-term, and long-term horizons, emphasizing interdisciplinary collaboration, scalable deployment, and governance.
References
Ahmed, S. Ozturk, and G. Mensah, "Bibliometric analysis of AI in education: Trends and geographic distribution," Educ. Inf. Technol., vol. 30, no. 2, pp. 1200–1220, 2025.
Aziz, A., Al Aydrus, A. T. A., & Astuti, N. V. (2026). Islamic Educational Leadership and Its Impact on Arabic Oral Proficiency. Leadership in Muslim Societies, 1(1), 75-92.
Birhane and H. Jobin, "Human-centered design and cognitive autonomy in AI educational tools," J. Learn. Sci., vol. 34, no. 1, pp. 88–117, 2025.
Bloom, Taxonomy of Educational Objectives: The Classification of Educational Goals. David McKay, New York, USA, 1956.
Braun and V. Clarke, "Using thematic analysis in psychology," Qual. Res. Psychol., vol. 3, no. 2, pp. 77–101, 2006.
Chen, T. Patel, and K. Nguyen, "Autonomous AI systems in higher education: Design principles and performance," IEEE Trans. Learn. Technol., vol. 18, no. 2, pp. 45–58, 2025.
Graesser et al., "Affective co-learning agents in intelligent tutoring systems," Int. J. Artif. Intell. Educ., vol. 35, no. 2, pp. 345–378, 2025.
Higgins et al., Cochrane Handbook for Systematic Reviews of Interventions, 2nd ed. Wiley, Chichester, UK, 2019.
Holmes, M. Bialik, and C. Fadel, Artificial Intelligence in Education: Promise and Implications for Teaching and Learning. The Center for Curriculum Redesign, Boston, MA, USA, 2023.
Imoize, D. Lee, and C. Tanner, "Security frameworks for cloud-based adaptive learning systems," IEEE Access, vol. 13, pp. 55230–55245, 2025.
JAnderson, C. F. Corbett, K. R. Koedinger, and R. Pelletier, "Cognitive tutors: Lessons learned," J. Learn. Sci., vol. 4, no. 2, pp. 167–207, 1995.
Joanna Briggs Institute, Critical Appraisal Tools. JBI, Adelaide, Australia, 2020.
Johnson and M. Lee, "Personalized learning pathways through agentic AI: A multi-site study," Comput. Educ., vol. 180, pp. 104430, 2025.
Khalil and N. Brouwer, "Longitudinal effects of adaptive learning on academic outcomes," Comput. Educ., vol. 182, pp. 104470, 2026.
Kim and J. Okonkwo, "Redefining teacher roles in AI-mediated classrooms," Teach. Teach. Educ., vol. 140, pp. 104320, 2025.
Lin and A. Park, "Agentic AI architectures: Memory, planning, and tool use in educational agents," Future Internet, vol. 17, no. 4, pp. 210, 2025.
Martinez and R. Faure, "Adaptive assessments calibrated to learner growth trajectories," Assess. Educ., vol. 32, no. 1, pp. 78–94, 2026.
Mikolov, I. Sutskever, K. Chen, G. Corrado, and J. Dean, "Distributed representations of words and phrases," in Proc. NIPS, pp. 3111–3119, 2013.
Nakamura, "Dynamic learning ecosystems: Lifelong learning agents and multimodal agency," Br. J. Educ. Technol., vol. 57, no. 3, pp. 678–695, 2025.
Romero and S. Ventura, "Educational data mining: A review of the state of the art," IEEE Trans. Syst. Man Cybern., vol. 40, no. 6, pp. 601–618, 2010.
S. J. d. Baker, "Data mining for education," Int. Encyclopedia Educ., vol. 7, pp. 112–118, 2010.
Torres and D. Singh, "Equity and accessibility in context-aware educational AI," Int. J. Educ. Res., vol. 125, pp. 102050, 2025.
Safitri, N. M., Hijazy, R. W., Mihrain, S. A., Septiadi, M. A., & Hadziq, M. F. (2026). Public Policy, Educational Quality, and Digital Culture: A Mixed-Methods Analysis of Indonesia’s Intellectual and Technological Gap. Tunjuk Ajar: Journal of Education and Culture, 2(1), 76-92
UNESCO, Recommendation on the Ethics of Artificial Intelligence. Paris: UNESCO, 2021.
Verma and A. Nkrumah, "Methodological quality in AI-education research: A meta-review," Syst. Rev. Educ. Res., vol. 12, no. 1, pp. 45–67, 2025.
Wells, B. Shea, and D. O'Connell, The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses. Ottawa Hospital Research Institute, Ottawa, Canada, 2013.
Williams, "Holistic student development with agentic AI companions," J. Educ. Technol. Soc., vol. 28, no. 1, pp. 12–25, 2025.
Zh, M. H. R., Sari, C. M., Thaariq, Z. Z. A., Sumalee, I., & Kamilah, A. N. (2026). Enhancing conceptual understanding in Islamic education through a hypercontent digital flipbook in blended learning: Evidence from Indonesian madrasah. Educational Technology in Developing Countries, 1(1), 1-22.
11.gif)
