If you have not been living under a rock for the past two years, you should be aware of the fact that AI poses significant challenges to assessment in the traditional philosophy classroom. By traditional philosophy classroom, I mean one in which students are expected to write philosophical papers of some sort based on the assigned readings in the course. In this paper, I outline why it is worth retaining this traditional practice of writing term papers as a chief means for student assessment and why we must intentionally limit the ability for students to utilize AI in them. If you are looking to be convinced of that, want more reason to be, or are interested in how to structure papers to evade the use of AI, I recommend reading it. However, the solution I outlined therein is uniquely fitting for in-person courses. After writing that piece, I found myself teaching from home with a newborn for two semesters. I then realized that the integrity of online philosophy courses poses a significantly more challenging problem (duh!)—one that no one seems to be doing much about. In this short piece, I would like to explore the nature of the AI problem for online philosophy pedagogy, the extent to which it is actually fixable, and some short-term solutions until larger institutional ones can be addressed.
The AI problem & online philosophy pedagogy
My guess is that a significant number of us are teaching online courses in which students are asked to complete discussion posts. Discussion posts have been the bread and butter of online formative assessment for a number of reasons: 1) they help gauge if students actually complete the reading 2) they help build some semblance of community 3) they stand in for in-class dialogue which is crucial for trying on philosophical ideas and concepts. I have yet to encounter an online philosophy course that does not include some form of the “discussion post” or kindred formative writing assessments in which students have to answer questions or reflective prompts on the reading. And, the unfortunate new reality that we are now faced with as educators is that nearly all of your students are surely using AI to complete them. Yes, most of them; I am sure of it. We encounter the same problem for argumentative essays. Because there is no way to control the classroom environment for AI use, you can be sure that a large majority of essays turned in for online courses are at least partly written by AI.
If this doesn’t bother you, it should. It means that we have no idea whether students know the content we teach online or if they can think philosophically without the use of a near-omniscient LLM as a co-reasoner. As a result, it means we have no idea if students have learned anything in our courses at all.
Though cheating was also uniquely easy for these assignments in online environments, it is now nearly impossible to prove and even more difficult to detect with any certainty. These realities make the temptation even more difficult for a poorly time-managed college freshman to resist. Thus, the sheer ubiquity of AI use for these assignments isn’t surprising.
Can we maintain the integrity of online philosophy courses?
In consideration of everything I’ve outlined above, I will take what I suspect to be a rather controversial position: we cannot maintain the integrity of writing assignments in online philosophy courses. This is a problem given that most online philosophy courses have writing assignments. Fundamentally re-thinking delivery and assessment structures for online classes (not just in philosophy but in the humanities generally) will be essential for preserving the integrity of our discipline and the utility of our courses in any core, undergraduate curriculum. If the word on campus is “take online philosophy, you can just use AI for all the discussion posts and papers and the instructors don’t care” we are in even more trouble justifying our existence than we were a few years ago.
Though we can still offer examinations and quizzes in a semi-controlled environment utilizing tools such as Honorlock, it is both soul-sucking, horribly tedious, and fails to assess a student’s true ability to do philosophy—that is, critically think through some philosophical problem in depth. I hate to be the bearer of even more bad news, but there are also AI tools students can download as plugins that help them cheat in online quizzes and when typing free-form answers in text boxes. So, what are we doing? What do we do?
Some short-term solutions
The suggestions I provide here are not perfect. Eventually, these kinds of solutions need to be coupled with larger institutional reforms for how online courses are designed and assessed. However, they do substantially mitigate the use of AI chatbot responses in assignments in the meantime. Most importantly, you can retain some joy in teaching philosophy online without becoming a full-time, online exam proctor.
- Annotations & Process Assignments
Ask students to complete hand-written annotations of the reading in place of discussion posts. They can do this by taking pictures of annotated, hard-copy texts or handwritten notes. My preference is the later and I have been calling these “Process Assignments” in which I ask students to demonstrate that they have a process for independently reading and trying to make sense of the material prior to coming to synchronous discussions.
There is also a wonderful application called Perusal that can be used to this same end, though Perusall is a social learning experience in which students can see comments from their classmates and respond to them. It requires some heavy lifting at the start of a new course but is a wonderful tool thereafter. I recommend checking it out here.
- Transcribed discussions in small groups
If your class meets synchronously, place students in small groups that meet weekly for a discussion using provided prompts on the assigned reading (given the day of to prevent students from preparing canned responses prior to the meeting). This can be at the tail end of scheduled class time or replace a live session all-together. They must turn in the transcript of their conversation and the corresponding video on CANVAS or whatever LMS you use. They get credit for participating but their participation must be substantive. The transcription makes it much easier to grade using a simple ctrl+F search for specific student’s names and contributions instead of watching videos of multiple discussions. There is also a way to use AI to count the frequency of participation from given students over time and the specific sort of responses you are expecting. However, that would require another post to explain in depth.
*I have to give credit to Simon Dutton for originally recommending a similar exercise that he utilized in online courses.
- Figma – Argument Map
Though I am thoroughly skeptical of assigning writing assessments in an online format, I do think asking students to map their arguments in detail prior to turning in their papers can decrease the likelihood that they will utilize AI chatbots to think for them as they write papers. For this to work, it needs to be more difficult to use AI on the map than to not use AI on the map. In other words, don’t ask them to turn in a Word document. Instead, introduce a tool called Figma (or something similar) that students must use to map their argument graphically. Figma is typically used by UE designers in tech. However, it has multiple functions and capabilities. It has a premade template for argument maps that you can find here. The tedium involved in copying and pasting bits and pieces of things from a chatbot into Figma may be enough to deter students from using them altogether. Ask them to submit a pdf of their map. They do not get credit for the assignment if Figma (or similar) is not used.
Lily Abadal
Lily Abadal is an Assistant Professor of Instruction at the University of South Florida, St. Petersburg. She is a virtue ethicist with interests in moral psychology, philosophy of religion, philosophical psychology, and pedagogy. Her work has been published throughout several edited volumes, the Journal of Religious Ethics, Teaching Philosophy, Religions, and Frontiers in Sociology. She has several forthcoming papers on AI and character formation in business and educational contexts.
