KOGI: A Seamless Integration of ChatGPT into Jupyter Environments for Program...
summary
KOGI embeds ChatGPT into Jupyter/Colab via the OpenAI API, captures code, errors and variable values, generates prompts and returns fixes or hints inline, and exposes a friendly AI-dog persona for Japanese classes. Deployed across two intro courses with about 150 students, it produces a measurable decrease in unresolved errors while 30% of students didn't use Ki in week one. Current limits: no output filtering, occasional breakage when upstream services change, and open concerns about overreliance and cheating.
AI-written summary · about
chapters (9)
- 0:00KOGI introduction
- 3:38Error message challenges
- 5:50AI-assisted education vision
- 9:27KOGI in action
- 12:24KOGI architecture and AI persona
- 16:26Classroom feedback
- 20:28KOGI functionality evolution
- 22:12AI in education insights
- 27:44Q&A
related talks
39mJupyter AI — Bringing Generative AI to Jupyter
46mAI Integration for Devs: ChatGPT and Copilot in Your Projects
32m"How Smart is AI? – Real projects to build real understanding"
38mSponsored Session: Jupyter AI v3: Human + AI Collaboration
45mSupercharge your coding: Integrating ChatGPT and Copilot in long-term projects
34mAI won't take your job but it sure will help you do it better
22mAI, Make It So - A Dive into the Universe of Prompt Engineering
12mUsing AI to Entice Employees to Python