Freely accessible generative artificial intelligence (GenAI) poses challenges to physiology education regarding learning and academic integrity. Although many studies have explored the capabilities of GenAI to complete assessments, few have implemented educative activities to highlight GenAI risks and benefits or explored physiology students' perceptions and uses of GenAI. Our study implemented a learning activity, designed using constructivist principles, to allow physiology students to explore GenAI and consider its use in assessment tasks. The activity engaged students (n = 236) enrolled in a second-year physiology subject over 2 years. The activity began with students being directed to critique a sample exam answer as a form of content revision. The answer had been covertly generated by ChatGPT, and it lacked depth and contained some hallucinated facts. Students then engaged in discussion about the use of GenAI for university study and assessment. Questions were used to stimulate thought and discussion, and student responses were collected via Padlet (492 posts). Thematic analysis of the comments highlighted students' beliefs about using GenAI and perceived benefits and risks. There was a general trend of increasing acceptance of using GenAI, and using it for assessment, over time. Students were concerned about breaching academic integrity guidelines, information accuracy and sources, and the negative effect it might have on their learning. At the conclusion of the activity, the revelation that ChatGPT wrote the sample exam answer reinforced the need for responsible GenAI use.