One of my favorite courses to teach is an upper-level anthropology seminar called “Animality and the Human Question.” Part multi-species ethnography, part critique of liberal humanism, and part philosophical debate on morality, we talk about everything from Hindu nationalism to the EU immigration crisis, wildlife conservation to industrial agriculture, scientific taxonomy to eugenic breeding under fascism, and Black civil rights to disability liberation. What I love the most is the course’s grounding in ethical conundrums that cannot be reconciled and political positions that are bound to be incongruous.

One of my teaching objectives is for students to develop skills engaging in high-level intellectual conversation on topics where there isn’t a clear-cut argument to make. In that spirit, the final requirement of the course is a twenty-five- to thirty-minute “oral exam.” I tell students that it’s meant to give them an opportunity to appreciate their intellectual growth over the term. What I don’t tell them is that it’s just as much about getting them to think and speak for themselves without the crutch that is ChatGPT.

Like many other instructors in our field, I have found the effect of generative artificial intelligence technologies on undergraduate learning devastating—nothing less than an existential threat to keeping the human at the center of the study of the human condition. I have tried several tactics to get my students to ease off the irresistible temptation to take shortcuts to their education: strict policies, grade penalties, signed contracts, “AI-proofed” prompts, and in-person and on-paper exams. I confess that I have yet to warm up to the idea of incorporating the technology into my course design (see Jenks et al. 2024), even as I have come to accept that the genie is very much out of the bottle.

I started doing oral exams in “Animality” before the generative AI floodgates opened, but I have only since come to appreciate how effective they are at encouraging the kinds of dialogic, collaborative, and synthetic thinking that no machine can replicate. After all, the only way to prepare for an oral exam, or so I tell my students, is to do the readings in earnest when they are due, to follow the arc of the lectures and discussions from week to week, and to pick up pointers from their classmates as we go along. Artificial intelligence cannot perform this work for them. Even if they were to try memorizing summaries or outlines produced by ChatGPT, none of that would be useful for the push-and-pull of a face-to-face conversation: “What do you mean by that?” “Which author are you drawing on there?” “I don’t think I agree with what you’re saying. Can you say more?”

Let me address the elephant in the room: Oral exams are scary. Some of my colleagues look at me with incredulity every time I mention that I use them in my teaching (“You do oral exams!?”). Faculty members and graduate students alike have told me that if they had been asked to do an oral exam as an undergrad, they would have “freaked out.” At best, these knee-jerk reactions are unsurprising given the close associations we have between viva voces and comprehensive qualifying exams for PhD candidacy. At worst, they are indicative of an academic culture where intellectual dialogue is presumed to be hostile.

I want to suggest that we revive the possibility of incorporating oral exams in anthropological pedagogy at the undergraduate level, particularly for small seminar classes of twenty-five students or less. Below, I offer a few strategies I have adopted to take the edge off what might otherwise seem to be an approach too intimidating to try out.

The first strategy is simple: I don’t use the words “oral exam” on my syllabi or instructional materials. Instead, I call it the “end-of-semester conversation” or the “EOS convo,” a term that, however awkward and unfamiliar, more faithfully captures the essence of the assignment. I mean “conversation” in the truest sense of the word: I ask students questions, and they ask me questions in return. I prompt them to stake a position on an issue, and they prompt me to stake mine. In an exam, learning is one-way; the student demonstrates an ability to answer the instructor’s questions. It is the kind of learning that AI is adept at mimicking. In a conversation, learning is two-way; student and instructor give and take from each other.

The second strategy, the very heart of the assignment, is to structure the EOS convo around three types of questions: (1) instructor-designed, (2) class-designed, and (3) individually designed questions. Not only does this tripartite structure provide students with the questions in advance, but it also enables them to play an active role in shaping them.

1. Instructor-designed

Two questions that I design myself and provide to the class three to four weeks in advance. In general, these are meta-cognitive questions that focus on the broad contours of the class, broaden the scope of inquiry to other facets of the discipline, and gauge how students’ thinking has transformed over time.

Examples

• We have spent the last several weeks talking about a pocket of humanistic and social scientific research loosely called “critical animal scholarship.” Some scholars have hailed this body of research as a progressive, social justice-oriented field. On the contrary, some argue that incorporating animals into serious scholarship is little more than a “trend with little substance.” The fact that nonhuman actors appear in theories regarding human rights, exploitation, and genocide, these critics contend, is ridiculous at best, and offensive at worst, cutting to the heart of anthropology as rightfully the “study of the human” (anthropos). Knowing what you do from this course, would you agree or disagree with these critiques, and why? (Hint: I can honestly see this going either way, so don’t think of this as a leading question).
• Is there a question—about any of the themes we’ve discussed—that you felt remained unanswered after our semester together?

2. Class-designed

Two questions that all members of the class design in dialogue with each other. The instructions I provide are: “What do you want me to ask you as a collective unit? What meaningful topics and themes do you think we should return to?”

Two weeks before the EOS convos are set to begin, I allot thirty to forty-five minutes of class time for students to discuss this amongst themselves. Organically, the session becomes an exercise in recalling the discussions that meant the most to them, teasing out the conceptual throughlines in their own words, and negotiating how to frame the questions democratically so that everyone has a way to do well (it’s still graded, after all).

As they deliberate, I sit to the side and pretend I am a fly on the wall. I am frequently impressed at how students choose to take on some of the most intellectually challenging themes when empowered in this way. The questions they come up with are often indistinguishable from ones I would have made myself.

Examples

• Over the course of the semester, we have learned to contend with the line where humans divide ourselves from animals. We have learned how unstable it is and how it is deeply embedded within (and drastically impacts) our politics, imaginations, and subjectivities. At the beginning of the semester, you were asked what you presumed to be the difference between humans and animals. How did you answer the question then, and how would you respond to the question now? Map out how your personal thinking has changed with the help of three to four key authors or ideas.
• We have analyzed a number of settings that have brought together humans and animals, from slaughterhouses to zoo enclosures to labs. Please reference at least two to three authors and/or texts to explain how “context,” broadly defined, informs the ways humans relate to, classify, or intersect with animal life? Why should context matter in the ways we strive for environmental protection, welfare, and rights?

3. Individually designed

One question that each student prepares for themselves to answer. The instructions are: “What do you want me to ask you, personally? What sort of question will allow you to showcase your individual strengths and insights gleaned from this course?”

I ask members of the class to turn these in about a week before the EOS convo, and I am always delighted to see how their questions reflect their intellectual preoccupations, curiosities, and commitments. See below for examples from former students, here anonymized by their initials.

 Examples

• B.A. (History, Pre-Law): What are the limitations of “animal rights” discourse? How can we think of and work towards collective liberation, for human and nonhuman beings alike, beyond the terms of “animal rights”?
• A.T. (Earth and Environmental Sciences/Anthropology): In what ways does anthropocentrism both limit and expand our understanding of boundaries across beings? How can we avoid situating the human as the standard, instead coalescing interbeing patterns across species to cultivate our mutual embodiment?
• C.C. (Philosophy, a dedicated vegan): Is the ultimate political project moral consistency and epistemological integrity? Is that even possible? Remember, from Blanchette (2020), it might not even be possible to type up this page and read it without touching traces of the industrial hog.
• E.O. (Education): How would you introduce someone entirely new to critical animal studies, or to the concept of humans and animality in general, to the conversation? What is the foundational knowledge they would need to know?
• J.E. (Art): One of the topics explored this semester has been that of the labor humans and nonhuman animals perform and the perceived value of said labor. Using two or three examples, how has this concept of labor and value been defined or challenged within the readings? In what ways have these definitions of labor and value, or challenges to labor and value, affected how we perceive the relationships between humans and nonhuman animals?

Finally, the third strategy is to have a grading rubric that is explicit about what students are being graded on and, critically, what they aren’t. In the few years of doing EOS convos, I have never once had a student who had done anything approximating an oral exam in their previous courses. Understandably, there’s a certain anxiety that goes along with not knowing what to expect. I have found that the rubric below, included on the syllabus from the start of the semester onwards, clarifies that my expectations aren’t unrealistic, even as they do set the bar high. Out of a 100-point scale, it might look something like this: 

• Mastery of the Material (20 points): The student demonstrates familiarity with a wide array of the course materials. They cite ethnographic examples, reference authors’ names, and draw on concepts from across the semester’s four units. They explain connections and disjunctions across the readings.
• Spontaneity (20 points): The student has internalized the materials’ arguments and converses freely about them without having to rely on a script. They can respond to the instructor’s follow-up questions, feed off the instructor’s comments, and ask their own questions in return.
• Critical Thinking (20 points): The student expresses their own opinions about the course materials and explains why they agree or disagree with a certain author’s point-of-view.
• Attentiveness and Engagement (20 points): The student incorporates examples from beyond the readings into the conversation. Importantly, they demonstrate that they have paid attention to the class discussions by citing examples or opinions that their classmates have raised.
• Reflexivity, a.k.a. the “Anthropological Perspective” (20 points): The student has reflected on how their assumptions have changed, shifted, or deepened over the course of the semester. They are able to articulate how they have challenged themselves to think differently than before, using perspectives gleaned from examples we’ve read from around the world, as well as from their classmates.

One thing that is not on the criteria, I make sure to emphasize, is self-confidence. I explain my reasoning in the following words, also included explicitly on the prompt: “Oral communication is an important skill for any engaged member of society. However, we all come from different cultural backgrounds that may or may not have helped us build up self-confidence at thinking aloud about difficult topics impromptu. We will incorporate a couple opportunities to practice beforehand. On the day of your end-of-sem conversation, if it takes you a few minutes to warm up and feel comfortable answering questions, that is totally okay. I will do my best to help put you at ease. What’s important to me is that your nerves do not get in the way of your ability to express your thoughts. Importantly, if you know that you will likely need to prepare to have this one-on-one interaction with me, feel free to set up an office hours appointment in advance.”

The main catch to the EOS convo is that it is a significant commitment of time, something that is only practical in the context of a small seminar. Even with my small seminar of twenty students, the convos take about ten hours total, which I divide across five days of the last week of class. Two hours, or four convos, is just about as much as I can muster for one day. That might sound like a lot, but thirty minutes per student is still less time than I normally spend commenting on research papers and figuring out whether AI was used to write them (an increasingly futile task). I allot five minutes at the end of each convo to jot down some notes and complete the grading. I don’t generally provide feedback beyond the kinds of comments that organically arise in conversation (“I really liked how you were thinking about x in y way.” “I hadn’t thought about z before.” “I think that engaging a, b, or c author a little more would have been helpful.”). I have never had a student contest their grade. That might be for the simple reason that most do very well, even those who I least expected would. If anything, I have found that the oral exam format provides me a platform to meet otherwise underperforming students where they’re at and to give them a chance at engaging the materials on their own terms.

In the context of a larger, lecture-based class, I can imagine an arrangement where graduate student teaching assistants (TAs) run the EOS convos in small groups of five and where the grading rubric is tailored to assess each student’s ability to engage their classmates. Depending on the size of the class, this may amount to less time spent providing feedback on paper drafts or grading exams with significant essay components. I have yet to try this out in one of my own lecture courses, but I know a colleague from outside of my department who has adopted the idea with success.

Above all, it’s worth mentioning how much fun I have engaging my students in a way that recognizes them as independent thinkers, so much so that the EOS convos are the one assignment of the semester that I grade effortlessly, even joyfully. “Oral exams” allow me to wrap up the semester with a clear idea about what, from the syllabus and from our discussions, mattered the most to students and why. They make it possible for me to impart the message that anthropological learning is about the process, and not the product. And that is something artificial intelligence cannot take away.