Generative AI as self-reflection mediator*

*Not yet attempted. Ideas based on elements and capabilities seen in conferences.

Description: Students use ChatGPT to complete a task and reflect upon it. Stage one sees students prompt the AI to act as a tutor and quiz them on the topic. Stage two sees the student provide a rubric to the AI to mark and provide feedback. Students then assess themselves and the AI based on the feedback, querying further about anything generated that could be incorrect or false before submitting the entire conversation for low-stakes marking. 

Assessment types applicable: Low-stakes formative assessment to ensure students are keeping up with weekly content or sub-unit content.

Unit Context: Any

Learning Objectives: Students will be able to Identify gaps in their knowledge, see how an “expert” would reply to questions, interpret the validity of “expert” responses and reflect on how to improve their work. 

Which AI tools & why: ChatGPT can assume roles and utilise rubrics to provide feedback and marking.

How is task & AI structured/scaffolded: Weekly or sub-unit material is covered in own-time reading, videos, and real-time workshops. This task would assist students in revising content. Using the provided prompt, get ChatGPT to quiz the student on salient points of content (provided in the prompt). Students reply and complete their quiz with ChatGPT. Students next provide a rubric and ask to be marked and provided feedback. Finally, students reflect on the feedback, citing areas for their improvement and also identifying anything that ChatGPT might get wrong. 

What support/guidance was needed for students: Tutor prompt, rubric marking, and feedback prompt provided for copy-paste.