Students in this course will use AI tools (ChatGPT, Copilot, Claude, and similar assistants) to write code — that is the reality of research computing today. This course does not try to prevent that. Instead, it teaches the underlying logic of coding so that students can:
- Direct AI effectively — write prompts that reflect a correct mental model of the problem.
- Read and evaluate AI-generated code — catch the subtle errors an AI model will produce, such as an incorrect data type, off-by-one indexing, silent propagation of missing values, or a misapplied statistical test.
- Recognize AI's limitations — know when a model is bluffing, hallucinating a package, or applying the wrong method.
- Own their analysis — a research paper is the student's responsibility, not the model's. The success criterion for this course is not “can you write code from scratch,” but rather: can you look at a block of AI-generated code and explain what it does, whether it is correct, and how you would fix it if it isn't.