Knowing Is Not Enough: Perceived Familiarity with Generative AI, Support Needs, and Critical–Responsible Orientation Among Higher Education Students

Generative AI tools are spreading rapidly through higher education, yet relationships among students’ perceived familiarity with GenAI, self-reported use, support needs, and critical–responsible orientation remain insufficiently understood. We surveyed 449 students from four Slovenian faculties of education and arts. Perceived familiarity with GenAI was measured using a single self-assessment item rather than a test of knowledge or competence. Critical–responsible orientation was operationalized as the importance students assigned to nine ethical and societal issues; it did not measure observed responsible behavior. Perceived familiarity with GenAI was positively but modestly associated with the frequency, breadth, and diversity of self-reported learning-related GenAI use. Descriptively, students with lower perceived familiarity reported greater support needs across all five areas, while responsible-use guidance was widely requested in both groups. Perceived familiarity with GenAI did not significantly predict critical–responsible orientation; age showed a small positive association. Among the 196 respondents who reported using GenAI for learning, perceived familiarity was the only significant predictor of the number of learning-related purposes; perceived instructional and institutional support and critical–responsible orientation were not independently associated with the outcome. These findings distinguish perceived familiarity with GenAI, self-reported use, support needs, and critical–responsible orientation as related but non-equivalent dimensions of GenAI readiness. Higher education institutions should provide curriculum-based support that combines practical GenAI skills with critical evaluation and attention to responsible use.