The integration of Multi-Agent Systems (MAS) with Generative Artificial Intelligence (GAI) is revolutionizing educational technology, creating unprecedented opportunities for personalized learning experiences that adapt to individual student needs in real-time. This emerging paradigm enables sophisticated educational ecosystems where specialized AI agents collaborate to provide comprehensive support across multiple learning dimensions, from automated feedback generation to intelligent tutoring and psychosocial assistance.

This systematic literature mapping analyzes the rapidly expanding intersection between intelligent MAS and GAI in educational contexts, providing educators, researchers, and technology developers with essential insights into current applications and future possibilities. Following established systematic mapping guidelines, we analyzed 16 primary studies published between 2020 and 2025, retrieved from major academic databases including Scopus, Web of Science, IEEE Xplore, and ACM Digital Library.

Our findings reveal remarkable exponential growth in this field, with publications increasing dramatically from 2024 onwards, indicating that educational institutions worldwide are recognizing the transformative potential of these technologies. Multi-agent architectures utilizing Large Language Models (LLMs) have become the predominant approach for addressing complex educational challenges that traditional single-agent systems cannot effectively handle.

The identified systems showcase diverse practical applications directly benefiting educators and students: automated personalized feedback systems providing instant, contextually relevant responses; intelligent tutoring agents adapting teaching strategies based on individual learning patterns; collaborative learning facilitators enhancing group dynamics; and comprehensive assessment tools offering evaluation beyond traditional metrics. Specialized agent roles include content generators, learning validators, pedagogical orchestrators, and student behavior simulators, demonstrating remarkable flexibility and scalability.

For educational practitioners, these findings have immediate implications. The technology enables learning environments supporting hundreds of students with personalized attention, provides 24/7 learning support availability, and offers consistent quality across different educational contexts. Systems can adapt to various learning styles, provide multilingual support, and maintain detailed learning analytics to inform pedagogical decisions.

However, critical challenges require educational community attention. Most implementations remain at proof-of-concept stages with limited robust empirical validation in authentic classroom environments. Ethical considerations regarding student data privacy, algorithmic transparency, and balancing AI assistance with human pedagogical expertise need urgent attention. The field lacks comprehensive methodological frameworks guiding systematic adoption and integration of these technologies.

This mapping provides the educational technology community with structured foundations for understanding and leveraging MAS-GAI convergence, emphasizing both remarkable potential for transforming educational delivery and critical importance of addressing implementation challenges to ensure innovations truly enhance learning experiences.

Keywords: Multi-Agent Systems, Generative Artificial Intelligence, Educational Technology, Intelligent Tutoring, Systematic Literature Mapping, Personalized Learning.