Morgan Banville
Massachusetts Maritime Academy
This assignment asks undergraduate students to co-create a classroom policy about generative artificial intelligence (GenAI). Through empowering students to make decisions in the classroom, this collaborative approach contributes to fostering both a community and a better understanding of the ethical, social, environmental, and political impacts and consequences of GenAI.
Learning Goals
Original Assignment Context: The following assignment has been taught at least 12 times (four courses over three semesters) in undergraduate courses such as College Writing, Business Communication, and Technical Writing. This has been taught predominantly in-person; however, the assignment has also been adopted for online Summer and Winter sessions. The class time is 50-minutes, however, depending on the specific class (such as my Surveillance and Professional Communication course), I usually split the information into two class periods, or one-and-a-half. The assignment has been very successful with students, as documented by their engagement, interest, and participation in the assignment. Students are curious, ask questions, and appreciate being asked to provide their own opinions and input.
Materials Needed
Time Frame: Two 50-minute class periods, or 1.5 class periods, usually completed during the first week of class
Overview: I have found that when discussing generative artificial intelligence with students, they are certainly using the tools but do not necessarily know 1. the implications of using GenAI (labor exploitation, intellectual property, and environmental concerns) 2. Ethical considerations (citational ethics, discrimination, inaccuracy, and more) 3. And how to navigate instructor policies. Because instructors often have their own syllabus statements about whether or not students can use GenAI in the classroom, it takes away the exploration of the tools and technologies in terms of student understanding. This assignment is an opportunity to co-create a class policy with student input, encouraging questioning, providing context to understanding both the affordances and issues with GenAI, and fostering student agency in the decision-making process.
Before the 50-minute class period, I assigned students to read: These Women Tried to Warn Us About AI, published by Rolling Stone. This article is a fantastic resource for students to be introduced to artificial intelligence and some of the concerns associated with the technologies. Then, at the beginning of the class, we start by exploring some of the takeaways. During this time, I also discuss with students how artificial intelligence is a big umbrella term that incorporates GenAI, which we will then spend the rest of the class discussing.
Depending on your class structure, student body, and lesson goals, instructors might also spend time reading in class (popcorn around) WTF is AI? | TechCrunch. After our class conversation about the Rolling Stone article, we then watch the following Vox video: AI can do your homework. Now what? This video provides a balanced discussion about ChatGPT and the pros and cons from high schoolers, college students, instructors, and more. As we watch, I ask students to jot down in what instances they believe ChatGPT, for example, could be used and instances that would be not allowed. It is important to note that we also discuss that there is many other generative AI other than ChatGPT!
Then, with students:
We work on the AI class policy synchronously in the Google doc, and brainstorm pros/cons. When there are instances of students suggesting copying and pasting their paper into the GenAI as I will note in the example, I usually turn to these articles:
The German press release highlighted copyright infringement concerns with pasting content, such as student essays, into GenAI (Dornis & Stober, 2024):
“As a closer look at the technology of generative AI models reveals, the training of such models is not a case of text and data mining. It is a case of copyright infringement – no exception applies under German and European copyright law,” says Prof. Dornis. Prof. Stober explains that “parts of the training data can be memorized in whole or in part by current generative models - LLMs and (latent) diffusion models - and can therefore be generated again with suitable prompts by end users and thus reproduced.”
This report suggests that the information students input in a generative model, is no longer their own intellectual property. In fact, such information has the potential to be generated as a response or output, depending on the prompt. This is concerning for our students, and something they should be aware of.
As we discuss each statement that the students write, I highlight, add commentary, and/or strikethrough the text. And this is decided as a class. Once we are done, the policy is then “locked” by being changed to “comment only,” rather than edit (students cannot see the comments if the permissions are set to “anyone with a link can edit”). Regardless, all information is kept, though, so we have a record of what we said is acceptable or not.
The following is an example from an upperclassmen undergraduate Technical Writing course. We created this document as a class within a Google Doc (since our campus is a Google campus). My own commentary will be italicized throughout the document.
To begin the assignment, I paste my initial statement that I include in the syllabus. The policy we create as a class then replaces the initial statement.
Initial Statement: A Note about AI and Chat GPT Writing
Technology, democracy, and writing are interwoven in our society. In this class, we rely heavily on technology to mediate our learning. With the recent advent of ChatGPT and other AI platforms that generate text, it is important to think about these technologies and how they can enter the classroom. I ask that you do not turn in work generated by AI generators as your own. We will talk about this in class and make a plan to address AI in the classroom together.
After the initial statement, I include a horizontal line between my statement and the student comments. After the line, I ask students: Can we use it [GenAI]? How and when? The bulleted list are student comments, unedited.
Can we use it? How and when?
After we discussed this specific bullet, I added a comment on the policy stating: We can use GenAI to brainstorm, but if an assignment asks for your own thoughts/opinions, don't use it! Also, be sure to fact check - the information might not be correct or accurate.
After adding this comment, we discuss as a class the accuracy and authenticity of the technology.
In regard to the previous bullet point, we had a class discussion about what it means to help or do the work. Helping might mean finding a topic (brainstorming), whereas doing the work is giving the “answer” or completing the assignment for you. Some students, especially those identifying as English as a Second Language Learners, discussed using GenAI for sentence structure. When this situation arose, we discussed intellectual property concerns and made sure to label entire papers that were pasted into GenAI systems as misuse.
In our class document, I used the strikethrough feature to cross out what we, as a class, decided was not an appropriate use of generative artificial intelligence. For example, the student comment about AI being used as "a tool for classes especially when doing work late at night and you're out of options and you still need help” is not an appropriate use. At this time, we also discuss topics such as academic integrity, ethos of GenAI, and my policies regarding emailing me about extensions on assignments.
Next, we discuss what is considered misuse in the classroom. In particular, we highlighted turning work completely generated by AI, as this would be considered an academic integrity violation.
What is misuse?
Lastly, we discuss as a class what happens or what the consequences will be if students misuse GenAI based on our class agreement (or policy).
What happens/consequences:
First time = zero
If it happens again, fail the class
Other option: get kicked out
Together, we agreed that the first misusage would be a zero. After the first misuse, a student would likely need to withdraw or fail the course. The last option, or the third instance of misuse, students suggested was to be sent to the Academic Integrity Hearing Board, and likely be suspended from the institution.
A challenge of this assignment is balancing the tension between what my own feelings are as an instructor towards usage of GenAI, and what students value. Because I value a space where students can contribute and collaborate towards the learning environment, part of the challenge is navigating such tensions. As you might notice in the student comments, there are some uses that I allow, despite my own approach towards refusing GenAI.
I want to acknowledge some of the sample resources that have been helpful in informing this assignment: