Survey of student perceptions and use of artificial intelligence when reading primary literature in the biological sciences

Artificial intelligence (AI) tools are increasingly used by students in higher education, yet little is known about how they are applied when reading primary scientific literature. This study investigated how undergraduate biology students use AI while engaging with primary research articles and how such use influences confidence, engagement, and perceived learning. Using a mixed-methods design across two institutions, we collected pre-course surveys assessing prior AI experience and attitudes, weekly self-reported AI usage logs during a semester-long primary literature-based course, and post-course surveys evaluating changes in practices and perceptions. Students entered the course with broad familiarity with generative AI but expressed uncertainty and concern regarding accuracy, ethics, and overreliance. Throughout the semester, students primarily used AI to clarify terminology, methods, figures, and background concepts, with usage patterns evolving toward more targeted and sophisticated prompting. Students also reported increasing awareness of AI limitations and the need for critical evaluation of AI-generated responses. Post-course results indicated increased confidence in analyzing primary data and deeper engagement with the literature, while concerns about overreliance largely diminished. Overall, students viewed AI as a supportive tool that complemented, rather than replaced, their own reasoning. These findings suggest that AI can enhance students' engagement with primary literature when used thoughtfully and underscore the importance of explicit instructional guidance to promote ethical, effective, and critical use of AI in biological science education.