PERC 2024 Abstract Detail Page
Previous Page | New Search | Browse All
| Abstract Title: | Exploring Large Language Models as Formative Feedback Tools in Physics |
|---|---|
| Abstract Type: | Contributed Poster Presentation |
| Abstract: | Significance of formative assessments and feedback is well-established in physics education, yet implementation in large enrollment physics courses poses substantial challenges such as scalability of timely personalized feedback. As part of efforts to productively incorporate large language models (LLMs) into physics education, we use a mixed method approach to compare human and artificial intelligence (AI) feedback to students on conceptual synthesis questions in an introductory mechanics course. We present our preliminary analysis showcasing the promising results and current limitations of tailored numerical and written AI feedback. We found that with physics instructors' guidance, AI provides relevant and timely written feedback to students. Nevertheless, AI struggles with edge cases and with specificity to students' answers, both of which are better handled by humans. Future work will investigate improving feedback quality by using rubrics to prompt AI, with the goal to enhance its potential utility to the physics education community. |
| Session Time: | Poster Session 2 |
| Poster Number: | B93 |
| Contributed Paper Record: | Contributed Paper Information |
| Contributed Paper Download: | Download Contributed Paper |
Author/Organizer Information | |
| Primary Contact: |
Shams El-Adawy Massachusetts Institute of Technology |
| Co-Author(s) and Co-Presenter(s) |
Aidan MacDonagh, Massachusetts Institute of Technology Mohamed Abdelhafez, Massachusetts Institute of Technology |
