Abstract: AI tools can now generate polished visualizations, scaffold Quarto reports, and debug data pipelines in seconds. So should statistics and data science students still learn to do these things themselves? The case for teaching programming and reproducible workflows is stronger than ever. When AI can produce code freely, mastery means being able to assess whether that code is correct, recognize when an analysis quietly goes wrong, and structure work so every step can be followed and verified. More importantly, learning modern data science workflows teaches students to think with data: to ask sharper questions, understand what data can and cannot answer, and reason rigorously about uncertainty and statistical models. The code is how we practice these habits; reproducibility is a cornerstone of scientific integrity and an essential skill for working productively with AI. Drawing on experience designing introductory data science courses and curricula, this talk reframes programming instruction in the AI era as a pathway to statistical thinking rather than syntax acquisition. Through concrete examples and ideas for assignments and assessments, it considers how to reward genuine understanding when a plausible-looking answer is only a prompt away. The goal is not to keep AI out of the classroom, but to prepare graduates to critically evaluate what it produces.

Mine Cetinkaya-Rundel Mine Çetinkaya-Rundel, professor of the practice and the director of undergraduate studies in Statistical Science, has been appointed director of the First-Year Experience in Trinity College of Arts & Sciences, Duke University.