It could be that you have used the Mapping your GenAI & AI Policy Tool and determined that student use of GenAI is fully permitted. This may be because you have redesigned or secured your assessments for the age of AI so that GenAI use won't undermine the learning outcomes. Or, perhaps you've redesigned them to incorporate GenAI use to amplify student achievement of the learning outcomes. Either way, you're going to allow students to use GenAI in your class, at least on the unproctored assessments.
In that case, you need a Full Use Policy that informs students how they can use GenAI in the completion of their academic work.
Craft a written policy that informs students they may use GenAI in your class. You can use/adapt this template.
Explain how and why students are allowed to use GenAI tools
Make sure to tie back to learning outcomes for the course, but also each particular assessment
Ensure students know that they are responsible for what they submit, whether content was generated by GenAI or not
So any inaccuracies, biases, mis/disinformation are the responsibility of the human author
NOTE: There are many ethical Concerns with GenAI: privacy, data security (FTC investigates OpenAI over data leak and ChatGPT’s inaccuracy), equity, biases, and inaccurate information. Because of these concerns, instructors shouldn’t require the use of GenAI tools but rather provide alternative options.
If you are not sufficiently familiar with GenAI tools, how they work, and how to best interact with them, then you can't teach students how to interact with them. And, if you can't teach students how to effectively use them, then we recommend not adopting a full use policy. The students we've seen in the UC San Diego Academic Integrity Office have no to limited understanding of these tools and definitely do not have a good sense of how to use them ethically, appropriately or properly. Create at least one activity that prepares students to use the tools properly and critically.
Teach Students High Quality Prompting. There are many models/frameworks for how to do this. We have found the following works best:
Role (e.g., "you are expert copy editor, experienced at working with college student to improve their writing")
Task (e.g., "I am going to give you a piece of my writing, and you are going to provide expert editing suggestions to make my writing more clear, concise and engaging")
Context (e.g., "I am a college sophomore majoring in MATH but writing a paper for my writing class. I am an average writer. The edits you provide should be reasonable for someone at my level and sound like me so your edits will be restrained.")
Steps (e.g., "first I will give you my draft paper and the grading rubric for the paper. Then you will evaluate it based on the rubric, telling me how you would score it. You will not edit my paper for me but wait for me to ask you for more information because I want to try to edit it myself based on your rubric feedback.")
Questions (e.g., "before we begin with your task, do you have any questions that would help you complete the task to my satisfaction?")
You can see that there is no fancy acronym for our framework (like those you'll find on the internet). But, the steps work.
Show Students how to Critique the Output
Allow students to submit one of the assessment prompts for your class to ChatGPT (for example) and have them critique the output using the grading rubric
Have students critically analyze the tools and their limitations
Give students two short essays - one written by a human and another written by a machine - and have them identify which is which and why
You might find this "Evaluating ChatGPT Generated Content" chapter from a free online book helpful
Require students take an AI Literacy course. You can create one yourself, or use one already available.
You always want students to document their use of GenAI tools, not only to protect academic integrity and help you assess whether learning occurred, but to raise their AI Literacy as well as their awareness of what they are doing and why.
See the Modeling and Encouraging Transparency page for more information.