Abstract:

About two decades ago, ePortfolios emerged as newly anticipated ICT platforms designed to complement or replace traditional Learning Management Systems (LMS), reflecting the shift toward learner-centered and outcomes-based education. Within this paradigm, ePortfolios serve as a foundation for reflective learning, enabling learners to review their achievements and understand their learning trajectories from a bird’s-eye perspective. However, despite these conceptual advantages, their practical adoption has frequently imposed substantial cognitive and administrative burdens on both learners and instructors.

A central challenge lies in the “selection and mapping” process, whereby learners must identify appropriate artifacts and align them with institutional competencies. Although essential for fostering reflection, this process is often perceived as stressful, leading to reduced engagement. Previous technological approaches, particularly the automated migration of data from LMSs, have largely failed to address this challenge. These systems primarily focused on reducing the effort required for data transfer rather than supporting the meaningful structuring and interpretation of learning evidence. Consequently, they often produced large volumes of unstructured data with limited pedagogical value.

To address these challenges, this study explores a personalized learning environment that integrates Generative AI. The proposed system aims to reduce learner burden by delegating high-load preliminary tasks, such as the initial selection and mapping of learning artifacts, to AI-based analysis. The system automatically analyzes submitted artifacts and generates a provisional mapping to a set of competencies independently developed for lifelong learning, drawing on established references such as OECD frameworks and the AAC&U VALUE rubrics. Although proprietary, this framework is adaptable to diverse institutional policies and diploma goals.

Furthermore, the system supports the intelligent integration of learning outcomes into a matrix-type ePortfolio. Unlike conventional ePortfolios that primarily function as chronological records, the matrix-type structure organizes learning evidence along two axes, forming a rubric-like grid. This design enables learners and instructors to visualize which competencies were developed through specific learning experiences.

However, this study acknowledges several critical limitations. First, the accuracy of AI-generated mappings remains a concern, as inaccurate classifications may distort a learner's self-perception. Second, learners may uncritically accept AI suggestions, potentially weakening the reflective processes ePortfolios are intended to support. Further refinement is required to ensure AI functions as a reflective scaffold rather than a substitute for learner judgment. Through the present work, we suggest that by facilitating the systematic integration of fragmented data, we can empower individuals to engage in continuous knowledge restructuring, therefore supporting sustainable growth throughout their lifelong learning.

Keywords:

Personalized learning environment, ePortfolio system, Individually optimized learning, Life Long Learning supports.