Beyond the final product: AI-Enabled Learner Traces as Assessment Evidence in Project Management Education

Abstract

Artificial intelligence (AI) is increasingly used in project management education (PME), especially in digital learning environments where learner activity can be captured as process data. This matters because project management learning involves collaboration, coordination, decision-making, reflection, and professional judgement, which are difficult to assess through final outputs alone. This systematic mapping review examines how AI-generated or AI-mediated learner data are collected, analysed, and used for assessment-related purposes, and identifies the educational implications reported in the literature. Following PRISMA 2020 guidelines, searches were conducted in Web of Science, Scopus, ERIC, and IEEE Xplore for studies published between January 2016 and November 2025. Nineteen studies met the inclusion criteria. The review identified five categories of AI-enabled learner data and assessment-relevant evidence: human–AI dialogue and prompt-based interaction logs, adaptive and personalised learning process data, project artefacts and revision traces, reflective, self-report, and course feedback evidence, and standardised task, output, and scoring data. These categories are treated as different layers in the data-to-evidence process rather than as learner traces in a narrow technical sense. Future-oriented conceptual data systems were identified as an emerging possibility but were treated separately from reported learner data and assessment evidence. Across the included studies, data were analysed using natural language processing, machine learning, rule-based techniques, learning analytics, and generative AI/LLM-based prompting, and were used to generate feedback, indicators, recommendations, and scoring support. Reported benefits mainly relate to formative feedback, improved visibility of learning processes, and more timely instructional support. However, the evidence base remains limited and uneven, with important gaps in validity, fairness, transparency, and longer-term educational effects. Overall, AI in PME is used not simply to automate assessment, but to broaden how learner activity is captured, interpreted, and used to support process-oriented assessment and pedagogical judgement.