KOGI: A Seamless Integration of ChatGPT into Jupyter Environments for Programming Education

Abstract

The impact of ChatGPT has brought both anxiety and anticipation to schools and universities. Exploring a positive method to improve programming skills with ChatGPT is a new and pressing challenge.
In pursuit of this goal, we have developed KOGI, a learning support system that integrates ChatGPT into the Jupyter environment. This paper demonstrates how KOGI enables students to receive timely advice from ChatGPT in response to errors and other questions they encounter.

We immediately introduced KOGI in our two introductory courses: Algorithms and Data Science. The introduction of KOGI resulted in a significant decrease in the number of unresolved student errors. In addition, we report on student trends observed in the classroom regarding the type and frequency of help requested. Although our findings are preliminary, they are informative for programming instructors interested in using ChatGPT.