Academic League of Artificial Intelligence - An Integrative Perspective of Teaching, Research, and Extension

The paper presents the organizational framework adopted by the Academic League of Artificial Intelligence (LIA) at the Federal University of Santa Catarina (UFSC), designed to integrate teaching, research, and university extension through a student-centered, project-based approach. The framework combines democratic governance, collaborative learning, and dynamic project organization to foster both technical and transversal competencies. The framework is illustrated through representative initiatives, including competition teams, study groups, open lectures, knowledge repositories, and AI-powered applications with social impact. These projects demonstrate how diverse educational, scientific, and extension activities can be developed within a common organizational structure while promoting leadership, scientific production, community engagement, and knowledge preservation. The reported experience indicates that the proposed framework provides a flexible and replicable model for integrating the three university pillars into engineering and computing education, offering practical guidance for academic leagues and similar student organizations.

The paper addresses the gap in literature regarding academic leagues in Computer Science and Engineering, noting that there is still very limited literature describing their adoption in Computer Science and Engineering and that little attention has been given to how these organizations can be structured to integrate the three university pillars while promoting student protagonism, technical excellence, and community engagement. The main contributions are threefold: "(i) we discuss the role of academic leagues as organizational structures for integrating teaching, research, and extension within Computer Science and Artificial Intelligence education. To the best of our knowledge, this is the first paper to examine academic leagues as an organizational framework for student-centered education in these fields; (ii) we present the organizational framework adopted by LIA, emphasizing student protagonism, project organization, and the articulation of the three university pillars; and (iii) we report representative projects developed within the league, discussing the educational outcomes, extension products, and lessons learned throughout its implementation."

The proposed framework is centered on the premise that students should occupy a leading role in both the conception and execution of projects, while faculty members act primarily as mentors and facilitators. The organizational structure combines a stable administrative structure with a dynamic portfolio of projects, where the administrative board is composed of students with roles including President, Vice-President, Secretary, and Communication Coordinator, elected annually through a democratic process. Projects have a flexible organization, each coordinated by a student project leader elected by participating members. The project development process involves projects originating from various sources, approval by the administrative board, creation of a call for participants, election of a project leader, and formation of study groups to investigate the scientific literature and state-of-the-art approaches. As projects mature, their initiatives are transformed into educational, scientific, and extension outputs, including technical reports, scientific publications, software artifacts, tutorials, online courses, workshops, public talks, or open repositories.

The paper details five representative projects. First, the LI(A)RA competition team, dedicated to developing intelligent agents for the Multi-Agent Programming Contest (MAPC), which originated from an undergraduate course, integrated study groups, achieved fourth place in the final competition, and produced a booklet, a self-instructional online course, and a book chapter. Second, the Open Talks about Artificial Intelligence, an annual cycle of lectures open to the broader community, integrated with courses, featuring student and faculty presentations, and permanently available on YouTube. Third, the Repository of Knowledge in Artificial Intelligence (RepoAI), an open repository containing tutorials, implementation examples, and educational materials, serving as the primary institutional repository for preserving knowledge and supporting onboarding of new members. Fourth, Study Groups, which are not independent projects but rather a collaborative learning strategy that supports every activity developed within LIA, providing the environment for acquiring theoretical background and often representing the first stage of the research process within LIA. Fifth, the GAIA (Games and Artificial Intelligence Applications) project, an extension initiative developing AI-powered interactive applications with social impact, including Forca Libras, a game using computer vision to recognize Brazilian Sign Language (LIBRAS) gestures, published at the Brazilian Symposium on Collaborative Systems (SBSC) 2026. Additionally, other extension initiatives include Visitas Guiadas (guided visits) for elementary and high school students and the Offensive Security in Practice (OffSec) project.

The lessons learned highlight that student protagonism as the driving force is a defining characteristic, with students progressively assuming greater responsibilities and developing transversal competencies. The framework integrates teaching, research, and extension through projects, relying on a complementary portfolio of initiatives whose combined contributions operationalize the inseparability of the three university pillars. Knowledge management and organizational sustainability are addressed by transforming project outcomes into persistent educational artifacts, establishing a cumulative learning process in which projects no longer represent isolated experiences but instead become part of the institutional memory of the league. The framework is replicable, as its principles are largely independent of the technical domain and none of the representative projects required large financial investments or specialized infrastructure. Limitations include that the reported experiences originate from a single academic league at a single institution, and the evaluation is primarily qualitative. Future work includes "evaluating the framework across multiple institutions and academic leagues, as well as conducting quantitative studies to assess its impact on student learning, leadership development, scientific production, retention, and engagement in university extension activities."

Improvements to AI Systems:

1. Student-Centered AI Project Governance Module

• Implement a framework where AI systems can autonomously propose, organize, and execute projects with a democratic decision-making layer (e.g., voting among AI agents or user stakeholders) rather than top-down control.

• Enable AI to rotate leadership roles among agents or users, mimicking the elected student board, to distribute responsibility and prevent single-point bias.

2. AI-Driven Knowledge Preservation and Onboarding

• Build a self-updating knowledge repository (like RepoAI) where AI automatically generates tutorials, code examples, and documentation from completed projects, then uses these artifacts to onboard new users or agents via personalized study paths.

• Allow the AI to tag and link past project outcomes (e.g., papers, code, videos) into a cumulative memory, reducing redundant learning and enabling faster iteration on new tasks.

3. Transversal Competency Development via AI Mentorship

• Design an AI mentor that tracks user or agent progress across technical and soft skills (leadership, communication, collaboration) by analyzing project contributions, meeting participation, and peer feedback.

• The AI can then suggest role rotations or new project types to fill competency gaps, similar to how LIA encourages students to take on increasing responsibilities.

4. AI-Powered Extension and Social Impact Project Generator

• Create an AI system that identifies community needs (e.g., accessibility, education, health) and generates project proposals with clear social impact metrics, mirroring the GAIA project (e.g., sign language recognition).

• The AI can also manage the full lifecycle: recruit participants, form study groups, and produce deliverables like open-source software, tutorials, or public talks.

5. Replicable AI Framework Across Domains

• Develop a meta-framework that abstracts the LIA organizational principles (democratic governance, project portfolio, study groups) into a configurable AI system, allowing it to be deployed in different fields (e.g., healthcare, climate science) without requiring domain-specific retraining.

• The AI would automatically adapt its project approval, knowledge management, and mentorship functions based on the target domain’s constraints.

What the Improved AI System Can Do:

• Autonomously run a student-led research lab: Propose projects, elect virtual leaders, form study groups, and produce publishable outputs (papers, code, tutorials) while tracking member skill growth.

• Serve as a lifelong learning companion: Continuously preserve its own learning history, generate new educational materials, and guide users through progressive skill acquisition.

• Bridge academia and community: Detect real-world problems (e.g., language barriers, educational gaps) and deploy AI-powered solutions with measurable social impact, while documenting the process for replication.

• Self-evaluate and improve: Use qualitative and quantitative feedback (e.g., user engagement, project success rates) to refine its organizational strategies, similar to LIA’s future plans for multi-institutional evaluation.

• Operate with minimal resources: Function on standard hardware and open-source tools, making it accessible to underfunded institutions or individual learners, just as LIA’s projects required no specialized infrastructure.

Academic leagues have become important mechanisms for promoting extracurricular education and strengthening the integration between universities and society. This paper presents the organizational framework adopted by the Academic League of Artificial Intelligence (LIA) at the Federal University of Santa Catarina (UFSC), designed to integrate teaching, research, and university extension through a student-centered, project-based approach. The framework combines democratic governance, collaborative learning, and dynamic project organization to foster both technical and transversal competencies. The framework is illustrated through representative initiatives, including competition teams, study groups, open lectures, knowledge repositories, and AI-powered applications with social impact. These projects demonstrate how diverse educational, scientific, and extension activities can be developed within a common organizational structure while promoting leadership, scientific production, community engagement, and knowledge preservation. The reported experience indicates that the proposed framework provides a flexible and replicable model for integrating the three university pillars into engineering and computing education, offering practical guidance for academic leagues and similar student organizations.