AI in Data Science Education

Project Overview

Over the course of the Spring 2024 semester and continuing into the Fall 2024 semester, the Data Science and Analytics Master’s program at Georgetown University, lead by professors Purna Gamage and Jeff Jacobs, piloted a series of innovations incorporating Generative AI tools into the program’s course curriculum and into departmental seminars held for the benefit of students and the Georgetown community more broadly. 

In the updates to the DSAN 5100: Probabilistic Modeling and Statistical Computing curriculum, Generative AI tools played a significant role in enhancing practice-problem generation. The goal was to demonstrate how ChatGPT can serve as a self-assessment tool, enabling students to gauge their understanding of course topics and identify areas that may require further study. To achieve this, the team expanded sections of the course that previously offered premade practice problems to include prompts for generating additional practice problems with ChatGPT. These prompts allow students to customize the difficulty level of the problems, providing a more tailored learning experience.

Outcomes

The key benefit of the DSAN 5100 curriculum update is the ability to tailor study materials to meet students’ varying needs. Within a particular course topic, some students may only need to work through one or two practice problems before the concept ‘clicks,’ while others may require a dozen or more. As students move through different course topics, those who quickly grasped the previous concept may find they need to engage with a larger number of practice problems to master the next one. This curriculum is an exciting example of how Generative AI can enhance pedagogical effectiveness, making course content more personalized and dynamically adaptable to each student’s needs.

The priority has been to identify which specific aspects of the courses, seminars, and workshops could most benefit from the capabilities of Generative AI, while also recognizing the areas that still require human creativity and effort. Student learning can be enhanced through this “division of labor” between tasks amenable to Generative AI enhancement (such as practice-problem generation) and tasks for which human compassion and creativity remain crucial (such as one-on-one conversations with our students during office hours). The more that Generative AI tools can be utilized to enhance tasks of the first type, the more uniquely human abilities can be applied to tasks of the second type.

Team