NeoEduc: A Synchronous AI-Driven 3D Avatar Tutor for Middle-School Robotics

• José A. Sebastião UFRPE
• Lucas S. Figueiredo UFRPE

Resumo

Traditional education, particularly in hands-on subjects like robotics, often struggles to provide personalized, real-time student support due to teacher time constraints and students’ reluctance to seek help in a public setting. This paper introduces NeoEduc, an AI-driven tutoring platform that presents a 3D avatar for conversational guidance in a virtual robotics lab, allowing middle-school students to interact via voice or text and receive adaptive feedback tailored to their performance history. The solution integrates an adaptive large language model (LLM) back-end to offer scalable, engaging, and private pedagogical support. We evaluated the platform with middle school students (N 20) in a robotics learning environment, using standard usability and user experience questionnaires (SUS, UEQ-S, NPS), as well as qualitative analysis of interviews and observations. The results demonstrate high usability (mean SUS score of 78.0) and user satisfaction (NPS of +75). Qualitative findings revealed increases in student engagement and autonomy, a potential reduction in teacher workload, and a notable acceleration of task completion, particularly in a comparative experiment where the NeoEduc groups significantly outperformed the control groups. NeoEduc presents a framework for creating educational experiences that leverage AI and virtual environments. The system’s ability to be configured by teachers towards ethical and pedagogically aligned posture highlights its potential to scale adaptive and immersive learning while mitigating social pressures in the classroom.

Palavras-chave: Intelligent tutoring, Robotics education, Virtual avatars, Large language models

Referências

R. Arlindo Filho, R. Z. Fanucchi, I. Winkler, and A. G. de Siqueira. The generative AI revolution in immersive technologies. In Simposio de Realidade Virtual e Aumentada (SVR), pages 82–84. SBC, 2024.

K. A. Baffour and T. Oyekoya. Effect of look-alike avatars on students’ perceptions of teaching effectiveness. In Proc. IEEE Virtual Reality (VR), 2023.

P. Bassner, E. Frankford, and S. Krusche. Iris: An AI-driven virtual tutor for computer science education. In Proc. ACM on Innovation and Technology in Computer Science Education (ITiCSE), pages 1–8, 2024.

J. Brooke et al. SUS – A quick and dirty usability scale. Usability Evaluation in Industry, 189(194):4–7, 1996.

C. A. Busch et al. The disproportionate impact of fear of negative evaluation on first-generation college students, LGBTQ+ students, and students with disabilities in college science courses. CBE–Life Sciences Education, 22(3):ar31, 2023.

X. Chen and Y. Kim. Adaptive learning in AI agents for the metaverse: The ALMAA framework. Applied Sciences, 14(23):11410, 2024.

S. P. Chowdhury, V. Zouhar, and M. Sachan. AutoTutor meets large language models: A language model tutor with rich pedagogy and guardrails. In Proc. ACM on Learning@Scale, pages 1–11, 2024.

W. Gan, Z. Qi, J. Wu, and J. C.-W. Lin. Large language models in education: Vision and opportunities. In Proc. IEEE Int. Conf. on Big Data, pages 4776–4785, 2023.

E. Holzapfel. The Class of 2030 and Life-Ready Learning, 2018.

L. Izzouzi and A. Steed. The influence of avatar realism and similarity in shaping group dynamics in collaborative virtual environments. In Proc. IEEE Virtual Reality (VR), 2023.

L. Labadze, M. Grigolia, and L. Machaidze. Role of AI chatbots in education: Systematic literature review. International Journal of Educational Technology in Higher Education, 20(56), 2023.

J. Laine, T. Lindqvist, T. Korhonen, and K. Hakkarainen. Systematic review of intelligent tutoring systems for hard skills training in virtual reality environments. International Journal of Technology in Education and Science, 6(2):178–203, 2022.

B. Laugwitz, T. Held, and M. Schrepp. Construction and evaluation of a user experience questionnaire. In Symposium of the Austrian HCI and Usability Engineering Group, pages 63–76. Springer, 2008.

M.-Y. Liao, C.-Y. Sung, H.-C. Wang, and W.-C. Lin. Virtual classmates: Embodying historical learners’ messages as learning companions in a VR classroom through comment mapping. In Proc. IEEE Virtual Reality (VR), pages 163–171, 2019.

A. Lieb and T. Goel. Student interaction with NewtBot: An LLM-as-tutor chatbot for secondary physics education. In Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, pages 1–8, 2024.

Z. Liu, Z. Zhu, K. Yu, X. Ding, X. Li, Q. Ma, K. Chen, O. Poquet, and F. Li. ClassMeta: Designing interactive virtual classmates to promote VR classroom participation. In Proc. ACM CHI, pages 1–11, 2024.

A. Mayer, K. Kastner, J. Reichwald, J.-R. Chardonnet, and J. Ovtcharova. Comparing avatar and face-to-face collaboration in VR education: Concept and preliminary insights. In 2023 IEEE 2nd German Education Conference (GECon), pages 1–5, 2023.

Pew Research Center. How teachers manage their workload. Technical Report, Pew Research Center, April 2024. Accessed: June 14, 2025.

F. F. Reichheld. The one number you need to grow. Harvard Business Review, 81(12):46–55, 2003.

A. M. Ryan and P. R. Pintrich. Should I ask for help? The role of motivation and attitudes in adolescents’ help seeking in math class. Journal of Educational Psychology, 89(2):329–341, 1997.

M. Schrepp, A. Hinderks, et al. Design and evaluation of a short version of the user experience questionnaire (UEQ-S). 2017.

E. Thiaville, J.-M. Normand, J. Kenny, and A. Ventresque. Virtual avatars as children companions for a VR-based educational platform: How should they look like? In Proc. Eurographics Conference on Virtual Environments, 2020.