M11: Moving Forward with the Integration of Generative AI in Jupyter Notebook for Teaching and Assessment of Data Science.


By Kehinde Irabor


Information

Google Colab has been an easy alternative to the Jupyter notebook platform. It is free and easily available, with no need for local installations on your computer. Therefore, it has been a good tool for teaching and assessing data science with Python. In December 2023, generative AI option was added to Colab’s code cell. This is good for fostering learning programming and in-class activities. However, that also poses a problem with assessing students’ homework or projects. Students tend to over-rely on it and just type the homework or project question in the generative prompt. The class is an eight-week, undergraduate Spring 2024 class offered in an online asynchronous format with 12 students. The students were encouraged to attempt the coding first. If they are unable to complete it, they can generate with AI, cite it, and comment on how generating with AI helped with the unclear concept. The students’ comments would be evaluated. They would also complete an evaluation at the end of the semester. There would be a short virtual interview or oral exam on the final project at the end of the semester. For a large in-person class, the virtual interview can be replaced with an in-person test on a lock-down browser.