When AI Can Code, What Should Students Learn? (feat. Bernadette Boscoe)

“I think that a focus on mastering the basics without AI is where we should be, especially with undergraduate data science education.”

— Bernadette Boscoe

In this episode, we speak with Bernadette Boscoe of Southern Oregon University about her path through art, math, computer science, and information science, and how that background shapes her approach to data science. She discusses using CloudBank and JupyterHub to move students beyond their laptops, giving them experience with shared computing systems and the infrastructures they will encounter in research and professional work. She also reflects on unequal access to computing resources and the need for infrastructure designed around teaching, rather than simply giving students access to commercial platforms.

We also discuss machine learning reproducibility, the challenge of preserving a research group’s tacit knowledge, and her work on AquiLLM, an AI assistant that helps researchers navigate the knowledge and materials their teams have accumulated. Bernadette shares how students contribute to Green Crossing AI, a wildlife camera project near Oregon’s Siskiyou Pass, and what they gain from working with real research problems. She closes with her perspective on teaching in the age of AI: building strong foundations, developing critical thinking, and bringing computational thinking into K–12 education.

Read the full episode transcript

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit datascienceeducation.substack.com