Beyond the Instrumental Use of Artificial Intelligence: A Formative Experience for the Development of Critical Digital Competencies in Higher Education

Abstract

The growing presence of artificial intelligence in higher education has made its use increasingly common, but not necessarily more critical. Students and educators may know how to use AI tools while still lacking criteria for questioning their outputs, recognising their implications, or deciding when their use is appropriate. This study examines how Critical Digital Competencies for AI (CDC-AI) develop when a formative experience deliberately places these questions at the centre of AI use. A qualitative interpretive design was used with 19 participants from two higher education contexts: 14 undergraduate students and five postgraduate participants who were also teachers. Evidence was gathered through diagnostic and final questionnaires, reflective narratives, academic essays, and researcher observations, and analysed through iterative coding, comparison, triangulation, and conceptual refinement. What emerged was not a simple progression from limited to greater knowledge about AI. Instead, participants gradually connected four processes: critical understanding of AI, critical evaluation of information, recognition of ethical implications, and reasoned justification of decisions. Particularly revealing was the way these processes began to shape decisions about when, how, and for what purposes AI should be used in academic practice. Participants moved from valuing AI mainly for efficiency and task completion towards questioning outputs, considering possible consequences, comparing alternatives, and explaining their decisions. This movement led to the study’s main integrative finding: the construction of criteria for situated AI use. Rather than proposing an additional competency, the study interprets these criteria as the result of mobilising the four processes together when a particular situation requires judgment. On this basis, CDC-AI are understood as interpretative frameworks that can guide AI-mediated educational practice. Given the small, purposively selected sample, these findings are not intended for statistical generalisation, but they offer a basis for examining how critical judgment can be pedagogically developed in higher education.