Teaching Statistical Programming in the Age of AI

Abhishek Chakraborty

Introduction

  • Name: Abhishek Chakraborty (he/him), Associate Professor of Statistics

  • Affiliation: Lawrence University, Appleton, WI

  • Courses (all in R)

    • Data-Scientific Programming (205)
    • Statistical Machine Learning (208)
    • Statistical Modeling (255)
    • Advanced Data Computing (405)

  • I use AI for work and personal purposes, but I am still figuring out how to use it in my teaching.

AI Policy

Within this course, you are welcome to use AI tools (like ChatGPT, Gemini, Copilot, etc.) for your learning purposes at no penalty. However, you should note that the content generated by these tools may be inaccurate, incomplete, offensive, contain incorrect facts and fake citations, and may perpetuate biases. I encourage you to verify any content from these tools with a trustworthy source. Their usage may also hinder independent intellectual thinking and creativity. You will be responsible for any inaccurate, biased, offensive, or otherwise unethical content you submit regardless of whether it’s your work or comes from these tools. You may not submit any work generated by an AI program as your own. If you include such material, its contribution must be acknowledged and cited like any other reference material. Also, please remember to discuss (a) your prompt, (b) how you revised the AI’s output, (c) any follow-up questions you asked, and (d) what you learned from the exchange. AI is not a replacement for knowing and understanding the material. I encourage you to use these tools wisely and intelligently, aiming to assist you with the subject matter, support your learning and creativity, and produce better results efficiently. I am happy to discuss and clarify if you are unsure about the usage of these tools and what will be considered acceptable work in this course. You will not be allowed to use computers, phones, or AI tools during in-class quizzes and exams.

Ground Rules

  • We will treat AI as a permanent entity.

  • We are teachers, not cops.

  • We will focus on pedagogy over platform (R/Python or Copilot, Claude, Gemini, ChatGPT, etc.).

Discussion Prompts

I would like to focus our discussion on the following broad prompts:

  • Alterations in in-class teaching practices

  • Alterations in assessment practices

  • Observations about student work

In-Class Teaching Practices

  • I do live-coding without any AI use.

  • In 205, I taught dplyr verbs for 1.5 weeks, before introducing students to GitHub Copilot. Students had mixed reactions.

  • Helped a lot with syntax issues.

Questions to consider:

  • Did I trade a syntax problem for a critical thinking problem?

  • Seems like I could do more with AI, but am I creating an unstable foundation?

Assessment Practices

  • Mostly project-based in 205 and 405, final project (Kaggle prediction challenge) in 208.

  • 20-minute in-class, on-paper quizzes for 205.

  • Oral interviews for 208 and 405.

  • Ask students to write a AI-collaboration log.

  • Break up projects into phases - (1) no-AI phase, (2) AI-allowed phase, and (3) reflection phase.

Questions to consider:

  • What do I actually evaluate? Inline code comments, appropriate and professional visualization choices, etc.

  • How do I adjust my rubric?

  • Do I need to retire some assignments? What do I replace them with?

  • How do I challenge students to think and evaluate their coding practices critically?

  • How do I assess creativity?

Typical AI-collaboration log

In this course, we will treat AI as a collaborator, not a substitute for thought. This document serves as a log for your collaboration with AI tools (e.g., ChatGPT, Gemini, Claude, etc.) during Project 3. Please be specific in your descriptions. Failure to document AI usage or representing AI-generated work as entirely your own will be considered an honor code violation.

Mention the AI tool(s) you used:

  1. Idea Generation: Describe how you used AI to brainstorm your topic, find potential datasets, or refine your research questions. What were some typical prompts that you used? How did the AI’s output influence your project?

  2. Code Generation: Describe how AI assisted in writing, debugging, or optimizing your code. How did you verify and correct AI-generated code? Did the AI provide any incorrect or “hallucinated” code? What did you think were some benefits of AI-assisted coding for this project? What were some challenges?

  3. Working with the Write-up: Describe how AI assisted in your writing of the blog content. What did you think were some benefits of AI-assisted writing? What were some challenges?

  4. Final Reflection: In separate paragraphs, reflect on (a) how the use of AI changed your learning process for this project, and (b) your overall experience with this project (challenges/achievements/skills you learned, etc.).

Student Work

  • Students seem to be doing more (going beyond classwork) with self-reported more efficiency.

  • Common student comment - Has been helpful in generating ideas(?) or refer to datasets for open-ended projects.

  • Less attendance in student hours, less interactions about the course content.

  • I am concerned about their learning, and their ability to transfer their learning to a different context without AI.

Questions to consider:

  • What changes in student work have you noticed?