This piece shares three perspectives on AI from professors negotiating its use in the anthropology classroom.

Ted Maclin: Anthropology and Artificial Intelligence

Download Ted Maclin’s syllabus and AI policy here

My thinking on artificial intelligence continues to evolve. My current thought is that AI (specifically generative AI) is popular not because it does anything worthwhile, but because it fits easily into a set of cultural narratives about the moral value of progress, efficiency, and technology. Many of the obvious problems with AI have to do with training data bias, environmental injustices, lack of representation among programmers and software elites (who remain mostly male), and misuse of the AI toward unethical ends. All of these are real problems. In addition, I argue that the driving ideas of AI are rooted in standardization, quantification, and surveillance—to such a degree that even if we were to resolve all the other issues with bias, environment, and misuse, AI would remain fundamentally opposed to values rooted in situated knowledge, care, and embodiment.

As far as assignments I’ve started to lean heavily into student journaling, particularly in upper-level classes. For example, students in my 2000-level class Anthropology and Artificial Intelligence spend time every week (ideally daily) journaling on their encounters with AI in the field. Then every other week they write short reflections on their fieldnotes for discussion with other students. This also mirrors the process of taking fieldnotes during ethnographic work, so it does double duty by building this habit among anthropology undergrads.

Prompt to Students: For this class you are required to keep a private fieldnote journal. I like a paper journal; you may prefer a digital journal. It doesn’t matter but you should make it yours. Your journal is where you record what you notice, question, feel, and struggle with as you encounter AI in everyday life and in course materials. I do not collect or read your journal in full. Instead, you will periodically submit Fieldnote Memos (every other week), in which you select small excerpts (plural) from your journal and analyze them using concepts from the course. These memos show how you are using journaling as an intellectual tool to investigate your own entanglements with AI. Submit these in the discussion area.

For a fieldnote journal, I don’t give prompts but I do spend time early in the class giving examples of the kinds of things that I personally include in my fieldnotes, and I refer to the journaling process throughout the course. Students have responded positively to the journals—several students have explicitly noted that journaling has let them see changes in their own beliefs and perceptions over the duration of the course.

I increasingly avoid the kinds of assignments that AI supports (e.g. essay questions on online tests) which, in my view, are not terribly effective at helping students to develop an anthropological praxis anyway. I teach a wide range of classes, and in most of them my approach is to include time thinking about what “technology” means, how technology and culture interact, and how power and technology shape one another. I find that some version of this discussion fits very naturally into courses on human evolution and prehistory, community development, globalization, environmental justice, and more. I’ve come to think that deconstructing popular ideas of technology may be one of the most important things we can do as anthropology faculty.

Samuel Collins: AI World—The Anthropology of AI

As we begin the semester, I want to remind you of the importance of academic integrity. Part of what you learn here at Towson University is how to do your work authentically and honorably, and I am here to help you do that. In my course, you will be given assignments and I will provide you with specific instructions about how to complete them. Sometimes the instructions will tell you to do the work entirely on your own without consulting other people or material, including the Internet. At other times, you may be asked to work together or to use certain technology tools. You will never, in this class, be permitted to plagiarize, meaning to use the work of another person or a tool such as ChatGPT and misrepresent that content as your own, original work. In other words, if you are allowed to utilize generative AI, you can’t represent the AI as your own work—i.e., you need to cite it. If you are ever uncertain about how you should work on an assignment, please contact me.

This is an AI policy that I helped craft for our faculty center shortly after the introduction of ChatGPT 3.5 in 2022, and it is one that appears in my syllabus for my course on the anthropology of AI. I think it is honest, if ambiguous. I may have you utilize AI tools/I may not have you utilize AI tools. Hell—I may even ban the Internet! You’ll just have to wait and see what my fickle mind decides. 

In discussions on our campus, students have indicated a preference for something more definitive, and I can sympathize with that. With all of us bringing our own orientations towards the recent surge in generative AI, the confusion can seem like a trap for unwary students. And it does little to assuage students who refuse to use generative AI for a variety of principled reasons.

But anthropology is about engaging complexity. As Emily Bender and Alex Hanna remind us, “AI” is less a “thing” than a marketing category (Bender and Hanna 2025). Under that “AI” umbrella lie countless systems that might variously be called “AI.” For most of us, many of these toil under the surface—improving our photos, blocking spam, recommending music and films. Some of it is more intrusive in the form of AI-generated content and a dismal race to enshittification (Doctorow 2025). Some of it is more covert, but also more insidious in its impact on our everyday lives—black-boxed algorithms, Flock cameras. As I explain in my course syllabus: “You will not spend a single day untouched by AI, from recommender systems, biometrics, smart devices, credit systems, spam filters. And, since 2022, generative AI—which proliferates in the interstices of human-made content.” In other words, generative AI here is almost a distraction, while under the surface AI applications of all sorts determine employment, benefits, aim weaponry, and classify and sort us.

The role for anthropology here goes beyond allowing or forbidding generative AI; it includes an interrogation of our AI lives at every level, from the ways recommender systems structure our consumer choices to ubiquitous surveillance and the AI-enabled necro-state. Student work needs to engage that complexity as well. There’s no easy binary “yes/no” to the AI world in which we live, and acknowledging that complexity is one step in establishing critical alternatives. We may give students options in their use of generative AI—including the option to eschew these text-extruding applications in their own work. On the other hand, there’s no escaping generative AI in Google search engines (even if you hide it from your results), and, more and more, no way of opting out of chatbots that have been shoehorned into the everyday applications of scholarship. But acknowledging that shouldn’t devolve into melancholy inevitability. Even as AI and other algorithmic tools infiltrate every part of our social, political, and economic lives, alternative formations arise that suggest more just, more transparent, and more sustainable possibility. One way to these alternatives, though, lies with our capacity to trace the elaboration of AI itself. 

And, as constellations of AI systems, generative AI and agents continue to proliferate in higher education and in our lives, I will need to adopt different course policies and practices, but they will continue to confront the complexities of “AI World.”

Klāvs Sedlenieks: Challenges and Opportunities

As an associate professor at Riga Stradins University (Latvia), I teach anthropology to undergraduate media students and master’s social anthropology students. When I started teaching around the turn of the millennium, I was shocked how widespread copy-paste plagiarism was in Latvian universities. Lack of awareness as well as lack of knowledge (both among students and teachers) was part of the problem. I was actively involved in designing the learning environment in which students who study hard would be treated fairly, but those who did not would not pass through. That included the effort to assess the actual knowledge of the students. This background has shaped my attitude towards the challenges of AI.

With the arrival of the AI systems, virtually all the previous tools that we used to probe student skills have become mostly useless, as neither the text nor prepared presentations, nor even in some cases seminar discussions, reflect the engagement of the students with the subject. I realized that unless they actively don’t want to, students will outsource any obligatory task to the AI. As a teacher I might be then drawn into the hopeless attempts to figure out what is and what is not the result of their own effort. I still think that reading the original is crucial because in anthropology it is not only what is being said and argued but also how it is said that is important.

My attempt to solve this problem has been to eliminate all obligatory tasks and allow the students to engage with the material freely as much as they wish and by whatever means they want. The only condition: At the final exam they will need to demonstrate knowledge of what anthropology has to say on the questions the course deals with. On my side, I make myself available for all the guidance and mentorship they might require or wish to have.

My first step a couple of years ago was to get rid of essay writing and shift to group presentations, but I experienced students presenting AI-written material, resulting in some presenters not comprehending the phrases on the screen. Then I made a more radical step: Since last year my undergraduates have had a completely free course except the final oral exam. They are free to attend my lectures, they are free to come to my seminars, and they are invited to come to individual supervisions. If they decide that they would like to train their writing skills, I provide suggestions for essay topics. I include the exam questions in the course description and all the course is structured along these questions—each set of questions has a dedicated lecture and a seminar discussion. The students are completely left to their own motivation and there is nothing except for the final exam for them to fear. In principle, if they invest a moderate amount of attention and effort, they should be well prepared for the exam.

I am still learning how to do this and how to guide students. Deadlines and obligatory tasks traditionally have been a part of student experience, and it might be very challenging to do lengthy reading of complicated texts if there is no deadline to miss. Thus, some students tend to skip all classes altogether hoping that they might be able to read and understand everything on their own (some indeed can do that). Others would come to the seminars and just listen to what their colleagues say. Moreover, reading of texts now can be substituted by listening to an AI-generated podcast and then having some conversation with it about the text. Lack of essay writing strips the students of the opportunity to undergo the tutelage of writing and the thinking that goes into writing. For many of them this is the first oral exam in their lives. During the exam they draw one of the sets of questions and have about fifteen minutes to prepare the answer. I allow them to give a narrative about the questions and then ask additional questions probing how deep is their knowledge and ability to use anthropological concepts to interpret life around them. I have to say that many fail on the first attempt. Sometimes when the second is strikingly better, I ask what the difference was between the first and the second attempt, and in almost all cases the answer has been, “I actually read the texts.”

The master’s students have been so far highly motivated, and I have assumed that they would not present AI-written texts as their own. However, the reality is that some of them have done that, and I have felt awkward and insulted to mark and comment on something that is generated by an AI. So far this question is left unresolved, and I am still puzzled as to how to proceed. Writing and reading have been important tools for development of thinking skills over centuries; this can hardly be substituted with oral argumentation, and it is doubtful that it can be substituted with just good prompt design skills. We will still need to have exercises that are sufficiently challenging, which allow the students to get to know something and to develop their thinking, but what these might be is an open question.