Beyond technology acceptance: students’ perceptions and prompt engineering competence in higher education

The integration of Generative Artificial Intelligence (GenAI) into Higher Education requires understanding students' perceptions and competence in interacting with these technologies. This mixed-methods study examined perceptions of GenAI and documented prompt engineering competence in academic tasks among 89 Psychology and Biomedical Engineering students. Study 1 employed a 2 × 2 quasi-experimental factorial design comparing disciplinary area and instructional approach (guided vs. autonomous), complemented by thematic analysis. Students reported favourable views of GenAI, particularly for understanding, learning, guidance, and work planning, while emphasising the need to verify AI-generated information and improve reliability and contextual accuracy. No significant multivariate effects of disciplinary area, instructional approach, or their interaction were found. However, item-level analyses identified two items with higher scores under autonomous instruction after multiple-comparison adjustment. Study 2 examined documented prompt engineering competence in 24 group reports using five rubric dimensions and the exploratory Prompt Competence Index (PCI). No differences between instructional approaches remained significant after multiple-comparison adjustment, although descriptive differences in Iteration and PCI emerged among biomedical engineering reports. Overall, findings support contextualised, iterative, and critically supervised human–AI interaction, while suggesting that instructional differences are context- and indicator-specific. Further validation of the PCI in larger, diverse, and balanced samples is required.