GAI Policies

Summary: Use AI tools (Claude, ChatGPT, Co-Pilot, etc.) to help you in your learning, not to do the assignments for you. Treat them like another student: you can get them to help you, but you are responsible for your own work. Give them proper credit / attribution. Don’t claim their work as your own.

I encourage you to collaborate with AI tools to enhance your learning. I just lack the experience/expertise to know the boundaries of where they go from a tool that helps you learn more/better to one that allows you to avoid learning (intentionally or inadvertently).

Therefore, rather than a specific “policy” there is a general guideline: treat an AI tool like you would another student.

1. Collaboration is encouraged - except in the places where we disallow/discourage it.
2. Your assignments should come from you. (e.g., don’t turn in an AI’s work as your own). This is a fuzzy line.
3. You are responsible for your work. (e.g., if it gives you bad advice, don’t try to fall back on “an AI told me so”)
4. Give proper attribution. (e.g., if it helped you, say so)
5. When in doubt, ask the course staff.

Some examples:

1. It is OK to ask an AI to critique your visualization to get ideas on how to improve it. But, if we ask you to provide a critique of something, you should give us your opinion.
2. It is OK to ask an AI to fix your writing (e.g., spelling and grammar). It is a slippery slope from a spelling checker to having it come up with the answer. Even something like “the ideas and content come from you, not it” is hazy - use your judgment, and err on the side of giving attribution.
3. It is OK to ask an AI for help. I ask (the chatbot I am using right now) how to do things in Tableau all the time.
4. It is OK to get the AI to do (some of) the programming for you. You should be deciding what to program.
5. If we ask you to do something by hand (for example, a pen and paper design exercise), we really intend for you to think through it yourself.
6. You should do the reading and content surveys without AI tools. These have the intent of getting your opinions. We know that the agents have read everything and have opinions.

I encourage you to use AI tools as a collaborator - as you would another student. Think of the AI (Claude, ChatGPT, Gemini, etc.) as a fellow student who took the class last year. We know they can do the work. We know they’ve read the readings. They are a good consultant and can answer questions, or discuss interesting topics with you. But you wouldn’t turn in their work as your own - that’s plagiarism.

I allow (encourage?) you to use AI as an assistant. Have it fix your grammar, or write the programs for you. Ask it for feedback on your ideas.

The problem: where AI becomes a problem is in intent. The use of AI isn’t the problem: it is using it with the intent to avoid learning. Or really, doing the hard work required to learn.

Even as I write this, AI tools are capable of doing all of the assignments in this class. You can point Claude (and probably the others—Claude is the one I have access to and have been trying) at any of the assignments. Indeed, I am trying it - and often find it gives “over-achiever” answers.

There is no reason to waste our time grading an agent’s responses to the assignments - we know it can do a reasonable job. We want to hear your answer. Not just because we want to check what you know, but also because everyone has a different perspective on things. (I guess we could ask the agent to give us 75 highly varied answers to each assignment - I have actually done this for some assignments).

I thought I had this nailed. Then I listened to this podcast which made me realize: detecting AI use isn’t the problem. It’s detecting the student’s intent. I didn’t change the policy at all. I just changed how I think about it.

Penalties and Violations: I reserve the right to penalize people who abuse the policy.

This policy would be in the Limited GenAI Use: Support Tool Only category according to the MIT AI in Education Report. This report is one of the best things I’ve seen about AI in education.