(26) ChatGPT Assisted Development of Abstract and Poster Assignment in a Large Self-Care Online Course

Objective : To use the assistance of ChatGPT to create a course exercise and assessments for research and presentation skills development.

Methods: Faculty instructed ChatGPT to generate a list of "hot topics” associated with Over-the-Counter Drugs/Self-Care Products for students’ research. AI also created pre/post-project quiz questions to assess students’ topic knowledge, rubrics for abstract/poster evaluation, and a 10-week in-class project development timeline. Faculty edited generated content for comprehensibility. Students were unaware of AI assistance in project development.

Each of 45 small student groups selected a topic, narrowing it for development. Each group created an abstract, poster, narration, and provided group peer evaluations and feedback to other groups. Faculty utilized the rubrics and provided written feedback on abstracts and posters. Students reviewed all finished posters to increase self-care knowledge. The last class was utilized to present the 10 best faculty-chosen posters and to obtain student project feedback via survey.

Results: Approximately 15 of the original 45 ChatGPT developed topics and 10 of 45 quiz questions were eliminated; new topics and questions were generated as replacements. Four timeline descriptions were edited for clarity.

Pre/post-quiz data revealed no change in students’ topic knowledge (80.4/81.4%). Students reported the topics were interesting (99%) and sufficient (96%); instructions (96%), timeline (99%) and rubrics (96%) were clear; most time spent on the project was during class (81%).

Conclusions: ChatGPT was a helpful tool in creating first drafts for a complex assignment. It made the daunting task of creating rubrics/assessments for presentation skill development more manageable; however, clarity of prompt instructions and review/edits of generated content were essential. Since students narrowed down the topics, perhaps ChatGPT-generated pre-/post-quiz questions were too general and did not adequately measure knowledge change.