AI Generated Feedback for Students Assignment Submissions: A case study in generating feedback for students’ submissions using ChatGPT
Author
Term
4. term
Publication year
2023
Submitted on
2023-08-11
Pages
36
Abstract
In the evolving landscape of education, the transformative potential of AI is gaining momentum. While AI-powered systems contribute to teaching, a void persists in generating personalized feedback for assignments. This study bridges this gap by integrating AI-generated feedback within the UCN didactical framework, exploring a resource-efficient, student-centric approach. The study delves into prerequisites to empower a chatbot, utilizing a generative language model like ChatGPT, to craft tailored feedback for assignments. Educational context, assignment specifics, evaluation criteria, and an indicative solution serve as input prompts, generating valuable feedback aligned with lecturer perspectives. The analysis of AI-generated feedback for the assignments included in the case study reveals consistent quality. Developing a feedback system extends beyond ChatGPT, as reliability demands lecturer validation pre-student receipt, ensuring transparency. A pipeline leveraging the OpenAI API is proposed for operationalization. This efficient approach resonates with personalized feedback objectives, complementing pedagogical methods and prompting subsequent dialogues. The study's conclusion synthesizes findings and contemplates AI integration implications. Insights emerge on benefits and challenges, guiding future research. Establishing AI's role in feedback augments educational needs, fostering enriched student engagement.
In the evolving landscape of education, the transformative potential of AI is gaining momentum. While AI-powered systems contribute to teaching, a void persists in generating personalized feedback for assignments. This study bridges this gap by integrating AI-generated feedback within the UCN didactical framework, exploring a resource-efficient, student-centric approach. The study delves into prerequisites to empower a chatbot, utilizing a generative language model like ChatGPT, to craft tailored feedback for assignments. Educational context, assignment specifics, evaluation criteria, and an indicative solution serve as input prompts, generating valuable feedback aligned with lecturer perspectives. The analysis of AI-generated feedback for the assignments included in the case study reveals consistent quality. Developing a feedback system extends beyond ChatGPT, as reliability demands lecturer validation pre-student receipt, ensuring transparency. A pipeline leveraging the OpenAI API is proposed for operationalization. This efficient approach resonates with personalized feedback objectives, complementing pedagogical methods and prompting subsequent dialogues. The study's conclusion synthesizes findings and contemplates AI integration implications. Insights emerge on benefits and challenges, guiding future research. Establishing AI's role in feedback augments educational needs, fostering enriched student engagement.
Keywords
Documents