Artificial Intelligence Policy

Large languable models (LLMs) are here to stay, and they present a (somewhat still very new) reality for both course design and approaches to testing.

STAT 447 has a clear focus on programming methods as applicable to data science. Programming is one of the fields possibly best supported by the new artificial intelligence tools. It is understandably tempting to use these tools when faced with programming tasks and challenges---as you likely will in the course homework.

We consider the homework to be "open book" and generally allow all avaialble tools. In the past that have been books, or Google searches, or maybe queries on StackOverflow. But now all this is at the fingertips of the LLMs, and hence yours. We encourage you to experiment and find your best way of learning and progressing with programming. Homework questions are implemented via PrairieLearn giving you an opportunity to try different approaches. Eventually your mark will reflect your best effort.

Do not just optimize for best score by repeated prompts in LLMs. You need to be able to work without the LLM because that is exactly how exams in the computer-based testing facility (CBTF) work. Questions there are also implemented via PrairieLearn but the CBTF has no external network connectivity for you.

So while the homework is "open book", the tests correspond more closely to an individual oral. It is just you, and the questions. So your task during this course is to learn how to solve programming questions when you do not have access to an LLM. Use them to learn, use them to aid your understanding and thinking, but learn to be self-sufficient. Certain concepts have to be learned by practice. And that is what STAT 447 aims for. Use the tools when you can, but also be proficient when they are not at your fingertips.