If we are clear with what our learning objectives are, then it can definitely help us align what we deem as the appropriate use of AI.

For example of the learning objective is to understand what makes up a thesis statement and the learning goal/assessment is to develop our own thesis statements, then using AI to generate thesis statements for us would be a clear violation for this assignment as the students are supposed to be doing that themselves as part of the assessment/learning goal. However, if they are using AI to learn what makes a thesis better or worse than another, then it would be appropriate because they are learning through examples the components teachers might be looking for in a "good" thesis statement.

All this to say the policy might need to reference the learning goals/objectives as opposed to a specific "one size fits all" statement.

"Generative AI itself can be a learning tool– as anyone who gets into the tool and starts inputting queries and studying the output knows. Your synapses are firing as you write and read the rapidly generated text. It’s fun, and you’re likely wide awake, judging, speculating, disagreeing, agreeing, and doing all those things that happen when an engaged reader encounters text. These potentials can be used in the classroom."

Be aware of the limits of ChatGPT:

If you provide minimum effort prompts, you will get low quality results. You will need to refine your prompts in order to get good outcomes. This will take work.

Don’t trust anything it says. If it gives you a number or fact, assume it is wrong unless you either know the answer or can check in with another source. You will be responsible for any errors or omissions provided by the tool. It works best for topics you understand.

AI is a tool, but one that you need to acknowledge using. Please include a paragraph at the end of any assignment that uses AI explaining what you used the AI for and what prompts you used to get the results. Failure to do so is in violation of the academic honesty policies.

Be thoughtful about when this tool is useful. Don’t use it if it isn’t appropriate for the case or circumstance.

1)Using a product or technology (ChatGPT,  Dall-E, etc.) in the course to create or aggregate synthetic content for the course such as any course writing, brainstorming, generating a draft, a works cited source bank, answering discussion groups, substitute or impersonate you in zoom, classroom, breakout rooms, collaboration, group work, or other attendance activities, is considered academic misconduct similar to plagiarism or violations of the honor code.

2)Using a product or technology as identified above to help you edit your work (improve your quality as defined by course instruction, course definitions, standards & expectations, and rubrics) is not a violation of the honor code statement, as long as you cite the tool or identify it as a partner or augmentation in use for a course related activity before submitting your work.

Although large language models like GPT-4 are a lot less fun than Janet, they raise many of the same questions. Those of us spending most of our time teaching or learning (e.g., you and me!) have a lot to think about, and we will all be experimenting with different approaches before there is anything like a settled view of how AI should be used within an educational setting. This means you will likely need to navigate various—and sometimes conflicting—policies in your courses semester. This will be a pain, but I hope you will also see it as an opportunity to shape your own thinking about artificial intelligence and the regulations we may (or may not!) need to ensure our use is aligned with our deepest values.

In that spirit, the (tentative) policy for AI use in this course can be summarized in two sentences: You may not use AI if your use will make it harder for you to learn or harder for me to assess whether you have learned. All other uses are welcome and encouraged. Importantly, this does not mean you are free to use AI whenever you wish, or that you can use it if you don’t think it will be harmful. What, then, does it mean?

You are all welcome and encouraged to use AI to:

1. Explain the course material to you (i.e., help you better understand the concepts/arguments/relationships/processes you need to learn)

2. Demonstrate and/or model the skills you are learning (e.g., request examples of worked problems, critical questions, or well-written essays)

3. Self-assess your learning (e.g., ask for feedback on works-in-progress, ask to be quizzed on the subject matter, or ask to be engaged in a Socratic dialogue)

4. Complete activities that explicitly require AI use (e.g., activities designed to develop or assess your AI literacy)

With the exception of activities that explicitly require AI use, you may not use AI to:

1. Complete practice assignments. Most of the assignments in this course are opportunities for you to learn through practice. In these assignments, it is the process that matters and not the result. If you wanted to learn to swim, you might ask Janet to explain it to you, to show you how to swim, and maybe even provide feedback on your stroke. But you wouldn’t ask her to show up at the gym at 6 AM and swim your practice laps for you. Unlike the other support she is providing, this would actually make it harder for you to learn. In fact, practice is so essential to learning that it’s doubtful you would ever learn to swim at all.

Note that, among other things, this course aims to develop your ability to “read increasingly sophisticated texts,” and you will practice this skill by completing the assigned readings before each class period. This means you may not use AI summaries or explanations as a replacement for doing the actual reading. AI may be a useful pre- or post-reading aid, but understanding the substance of the text is not the same as understanding how to read that text.

2. Demonstrate that you have learned. Although most activities in this course are designed to help you learn, some are designed to assess what you’ve learned at specific points throughout the semester. In most cases, I’m interested in your independent development, so I need to determine whether your work, on its own, demonstrates growth. In these cases, I need to see what you can do without AI assistance. If I wanted to know how well you could swim, I would need to watch you swim, not Janet.

This is a lot to keep in mind this semester, so I will be sure to let you know when I want you to complete an assignment, or part of an assignment, without AI assistance. But as a general rule, remember that you will get the most out of AI if you think of it as a tutor, helping you to develop, rather than replace, your own thinking.

One last very important point: remember that the Janet we have has just been rebooted, and may give us a 🌵cactus 🌵at any moment. You should be prepared for these mistakes and verify all results before you rely on them. It is ultimately you, not Janet, who will be responsible for what you’ve learned.

********

[Note: I am not going to include the following language this semester because I’m still unsure how I would make it work logistically, but I include it here in case anyone wants to be more ambitious than me:

In some situations, negotiated with me in advance, students may also make targeted use of AI to overcome learning challenges peripheral to the goals of the assignment (e.g., second-language learners could use AI tools to translate exam questions into their native language). However, these accommodations do not extend to the knowledge, skills, or dispositions that are the primary focus of the assignment.]

Notice that teachers could use AI detecting software (whether we do or not is up to teacher discretion!)

To students:

"you take the lead" in shaping your final response

AI has inherent bias

AI lacks creative thinking skills

* AI tools have limitations (ie, they lack critical thinking to evaluate and reflect on criteria; they lack abductive reasoning to make judgments with incomplete information at hand)