Relying on models to make assessment decisions without a “human in the loop” (i.e., where the person who ultimately uses their own judgment to make decisions may use AI as an assistant) is inadvisable and violates OpenAI’s usage policies. Current models have biases and inaccuracies and cannot capture the full complexity of students or educational contexts.
Research conducted by others shows that ChatGPT can serve as an assistant to teachers by providing feedback to students. The model can offer another perspective on student work and help students think about and improve their drafts; teachers can use it to accelerate the provision of comprehensive, detailed feedback to students.
To gain a deeper understanding of the overall situation of the class, you can provide all students’ responses to ChatGPT and ask it to identify common themes and common areas of strength and weakness.