1. Introduction

Online schooling has evolved from an emergency substitute to a stable educational approach in many countries, with increasing support from AI-driven systems for tutoring, assessment, analytics and student support. While these tools offer personalisation and scalability, concerns have been raised about surveillance, bias, reduced teacher autonomy, and psychosocial risks. In their large-scale systematic review of 146 studies, Zawacki-Richter et al. (2019) highlighted that much AI-driven education (AIEd) research has historically focused on technical innovation rather than pedagogical or ethical implications, and that educators remain underrepresented in design processes. In contrast, Akgun and Greenhow (2021) focused on K-12 schooling, explicitly foregrounding ethical tensions such as data protection, fairness, and the developmental vulnerability of minors, and calling for normative frameworks adapted to school contexts. This article synthesises insights from these two traditions and situates them within the operational reality of an innovative international online school, where cross-border regulation, cultural diversity and student well-being are central concerns. Early AIEd systems were largely tool-oriented, comprising intelligent tutoring systems, automated grading and learning analytics dashboards. As documented by Zawacki-Richter et al. (2019), these systems prioritised technical feasibility over pedagogical and ethical integration, often marginalising educators in decision-making processes.
    Recent scholarship emphasizes that ethical governance must be embedded directly into the architecture of AI-supported online schooling rather than treated as an external regulatory layer. Systematic reviews in AIEd reveal that early deployments privileged technical optimization over pedagogical and ethical reflection, often marginalizing educators in design processes (Zawacki-Richter et al., 2019; Williamson & Eynon, 2020). In K–12 contexts, concerns surrounding privacy, fairness, learner vulnerability, and well-being have been foregrounded, prompting calls for transparent, human-in-the-loop systems (Akgun & Greenhow, 2021; UNESCO, 2021). Complementary work has highlighted the need to mitigate algorithmic bias and to ensure explainability in automated assessment and tutoring systems (Holmes et al., 2019; Yadav & Chakraborty, 2020).
    Ukrainian scholars similarly stress that large-scale digitalization in schooling requires principled ethical oversight, particularly regarding surveillance, equity, and data protection (Богатирьова, 2021; Шевчук, 2020). Studies on algorithmic grading and digital assistants further underline the necessity of transparency, proportional data use, and teacher mediation in distance learning environments (Коваленко & Панченко, 2022; Кириченко, 2023). Together with international conceptual frameworks advocating educator-centered AI ecosystems (Luckin et al., 2016; Zawacki-Richter et al., 2020), this body of literature supports the development of distributed human–AI support models in which personalized student agents, teacher-facing analytic systems, and human tutors jointly uphold psychological safety, inclusion, and pedagogical authority.
    However, recent advances in large language models (LLMs), multimodal AI, and agent-based architectures are enabling a transition towards more personalised, dialogical and role-specific AI agents. This shift is particularly relevant in K–12 online education, where ethical concerns such as developmental vulnerability, well-being, and fairness are amplified (Akgun & Greenhow, 2021). We argue that ethical challenges in online learning should be addressed by structuring AI roles, rather than by limiting AI capabilities. We propose that ethical challenges in online learning are best addressed not by limiting AI capabilities, but by structuring AI roles explicitly within a transparent human–AI division of labor.

References 

1. UNESCO. (2020). Policy Brief: Education during COVID-19 and beyond. United Nations. https://www.un.org/development/desa/dspd/wp-content/uploads/sites/22/2020/08/sg_policy_brief_covid-19_and_education_august_2020.pdf

2. Kapp, K. M., & Defelice, R. A. (2019). Microlearning: Short and Sweet. Association for Talent Development. ISBN: 9781949036732. https://www.td.org/product/book–microlearning-short-and-sweet/111915

3. Zimmerman, B. J., & Schunk, D. H. (Eds.). (2001). Self-Regulated Learning and Academic Achievement: Theoretical Perspectives (2nd ed.). Routledge. DOI: https://doi.org/10.4324/9781410601032

4. Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence Unleashed: An argument for AI in Education. Pearson. ISBN: 9780992424886. https://discovery.ucl.ac.uk/1475756/

5. Hodges, C. B., Moore, S. L., Lockee, B. B., Trust, T., & Bond, M. A. (2020). The Difference Between Emergency Remote Teaching and Online Learning. EDUCAUSE Review. https://er.educause.edu/articles/2020/3/the-difference-between-emergency-remote-teaching-and-online-learning

6. Akgun, S., & Greenhow, C. (2021). Artificial intelligence in education: addressing ethical challenges in K-12 settings. AI and Ethics, 1, 431–440. https://doi.org/10.1007/s43681-021-00096-7

7. Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review of research on artificial intelligence applications in higher education: Where are the educators? International Journal of Educational Technology in Higher Education, 16(39). https://doi.org/10.1186/s41239-019-0171-0

8. Holmes, W., Bialik, M., & Fadel, C. (2019). Artificial Intelligence in Education: Promises and Implications for Teaching and Learning. Boston: Center for Curriculum Redesign.

9. Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence Unleashed: An Argument for AI in Education. Pearson.

10. Williamson, B., & Eynon, R. (2020). Historical and social contexts of AI in education. AI & Society, 35, 611–620.

11. Zawacki-Richter, O., et al. (2020). Ethics and AI in Higher Education: A Theoretical Framework. Educational Technology Research and Development.