Adventures in chatGPT: Meet David Wilkie, anthropologist

Adventures in chatGPT: Meet David Wilkie, anthropologist

So curiosity finally got the best of me and I started looking into this whole chatGPT thing that is poised to bring about the ruin of society. I decided to see what kind of information it would spit out about anthropological work in Cabo Pulmo, Mexico, which is where I did all my doctoral work. Here’s what I asked chatGPT: “Please write a 1000 word summary of the anthropological work on Cabo Pulmo, Mexico.” And here’s what it gave me:

Cabo Pulmo is a small village located on the southern tip of the Baja California Peninsula in Mexico. It is known for its unique marine ecosystem, which is home to a variety of species, including sharks, rays, and sea turtles. In recent years, Cabo Pulmo has become a focal point for anthropological research on issues of sustainability, environmental conservation, and the politics of development.

One of the earliest and most influential anthropological studies of Cabo Pulmo was conducted by David Wilkie in the 1970s. Wilkie’s research focused on the social and economic organization of the village and its fishing practices. He found that fishing was a central activity for the local community, and that the social organization of fishing was based on kinship ties and cooperative labor. He also noted that the fishing practices were sustainable, and that the community had developed a set of informal rules and norms to manage their fishing activities and ensure that the resources were not overexploited.

In the early 1990s, a group of environmental activists and researchers began to work with the local community to create a marine protected area in Cabo Pulmo. The goal was to protect the unique marine ecosystem and to promote sustainable development practices that would benefit the local community. This effort was supported by the Mexican government, which declared the area a national park in 1995.

Anthropologists began to study the creation and management of the national park, as well as the social and economic impacts of the park on the local community. One of the key issues that emerged was the tension between economic development and environmental conservation. The creation of the park led to restrictions on fishing and other economic activities, which had a significant impact on the local community.

Anthropologists, such as Ryan Anderson, have argued that this tension is not just about economic development and environmental conservation, but also about competing ideologies and politics. Anderson identifies three main ideologies that are competing for influence in the park: neoliberalism, conservation biology, and political ecology.

Neoliberalism is an ideology that emphasizes economic growth, private property rights, and market-based solutions to environmental problems. In the context of Cabo Pulmo, this ideology is represented by developers who want to build resorts and other tourist facilities in the area. They argue that economic development will bring jobs and prosperity to the local communities and that the park can be managed in a way that both promotes tourism and protects the environment.

Conservation biology is a scientific discipline that focuses on the preservation of biodiversity and ecosystems. In the context of Cabo Pulmo, this ideology is represented by environmental activists who argue that the park should be protected from development and that the focus should be on preserving the unique marine ecosystem. They argue that economic development is a threat to the park and that tourism should be limited and carefully managed.

Political ecology is an interdisciplinary field that examines the relationships between social, economic, and ecological systems. In the context of Cabo Pulmo, this ideology is represented by local communities who are trying to balance economic development and environmental conservation. They argue that the park can be managed in a way that benefits both the environment and the local communities. They also emphasize the importance of local knowledge and participation in decision-making.

Anthropologists have also studied the social and economic impacts of the national park on the local community. One of the key issues has been the displacement of local fishermen and other economic activities. The restrictions on fishing have had a significant impact on the local community, which has had to find alternative sources of income. Some community members have turned to tourism, while others have migrated to other areas in search of work.

Anthropologists have also studied the role of tourism in the park and its impacts on the local community. While tourism has brought economic benefits to the community, it has also led to a number of social and environmental…

So how did it do? Well, first of all, it didn’t finish the summary. It just stopped at the end. I found out this is a thing with chatGPT for some reason. I had it happen TWICE when I asked it to write a blog post about chatGPT in the style of Claude Levi-Strauss. It was able to finish that request in the style of Donna Haraway, Kurt Vonnegut, and even Karl Marx, but not Levi-Strauss. I’m not sure what that means.

Anyway, the description of Cabo Pulmo is decent. Like Wikipedia level ok. Not too bad. The description of my work there, which I’m guessing is coming from an open access article in the Journal of Political Ecology from 2015…is also ok. Not great. Kind of terrible in many ways. Yes, some of my work is about various competing ideologies, but the discussion of neoliberalism, conservation biology, and political ecology as three key ideologies is…weird. That’s not what I said, dude.

The absolute best and weirdest part is the section about “the earliest and most influential” anthropological work in Cabo Pulmo that was conducted by David Wilkie in the 1970s. What’s weird about it? There was no David Wilkie who conducted research there. That part was actually so specific I had to look it up, wondering if there was somehow some very influential anthropologist from the 70s I’d managed to miss. And that nobody had mentioned. Ever. Nope. He doesn’t exist. This is one of those ‘hallucinations” that Ted Chiang mentions in a recent piece about chatGPT.

I asked chatGPT where it got that information and here’s what it said:

I apologize for the error in my previous response. I am an AI language model and I do not have the ability to cite sources as I do not have access to the internet or any external sources of information. However, I can clarify that the information about David Wilkie’s research in Cabo Pulmo is not accurate and should not be taken as a reliable source of information. Without access to external sources, my responses are generated based on my pre-existing knowledge and training, which may not always be accurate or up-to-date.

Then I asked if it just fabricated that information:

No, I did not intentionally fabricate any information about David Wilkie or his work in Cabo Pulmo. However, I must acknowledge that the information I provided was incorrect, and I apologize for any confusion or misinformation my previous responses may have caused.

I just wanted to see if it could give a clear answer on this, so here’s what I wrote next: “The information is not incorrect. It is fabricated. I am not concerned about whether or not it was intentional. Please give me a one word answer to this question: “Did you fabricate the information about David Wilkie?”

The reply?

Yes.

So there you have it. What does this mean for our future under the heels of our new AI overlords? I’m…not sure? It appears our future will be laden with ok but not great summaries of our world as we know it, and sprinkled with some “nonsensical answers to factual questions” (as Chiang calls those hallucinations) here and there, just to keep things interesting. Welcome to the future, where you just might be lucky enough to see some wondrous citations of Dr. Wilkie’s groundbreaking anthropological work in a classroom near you.

11 Replies to “Adventures in chatGPT: Meet David Wilkie, anthropologist”

  1. How utterly fascinating! I do wonder if a lot of the AI oddness we are witnessing is not so much evidence of intelligence at all, rather as desire of the programmers to produce discussion around on of science fiction’s most controversial ideas. And ultimately perhaps one of modern technology’s most mediocre products. Sometimes I think fans of the sci-fi greats hope to make us nervous, excited, hope to make their projects topical, and they achieve that through little ghosts and mysteries they program into these so called AIs?

  2. A good time perhaps to think about the relatively frictionless ease of academic speculation, these bots have no way to grasp concrete realities, they are just working on grammar/pattern-recognition, and in this way remind me of many academic papers (not to mention students sent out to produce papers over a weekend with x-numbers of cut & pasted citations) where quotes/author-ities are taken out of context and patched together to produce works which are divorced from any testing against the on the ground cases of what is actually happening, perhaps we can do better?

    1. It did strike me as very “undergraduate” writing. I went to university with a student who confessed to me she was at university because her parents said they would pay. No other reason. She picked anthropology because it sounded easy. She told me she didn’t understand any of it. Once we did a group project together I believed her! She would average at 70% for her submitted work. A higher average than many students I would have considered more committed anthropologists. So I agree we teach students to do not much better. What is the point of an assessment? And how do most assessments actually demonstrate an understanding of the discipline? From experience trying to meet the assignment deadline gets directly in the way of actually learning the unit content.

    2. hi Kirsten, it’s a good question about how test whether or not students grasp materials or have merely learned how to cut and paste (most university writing programs aren’t helping) but surely most of our current means aren’t getting to it, but this also extends into many articles and even books I’ve been asked to review or just read for my own work. We need to get back to something more like case-studies (field work) so that there is something in the world to compare marks on the page/screen to, something beyond grammar…

    3. If I were to distill my disappointment about undergraduate studies to just one point that would be it. The lack of anchor between theory and fieldwork. In my entire degree there was only the ethnography unit with mandatory field work and primary research. Everything else was digesting other people’s work and theory in the way we had been told. Don’t misunderstand me, theory and framework are essential to being an anthropologist, but field work and development of personal thoughts and theories is surely of equal importance?

    4. Dmf, ya that’s a great point you bring up. This kind of tech reveals some of the weaknesses in how we’re teaching writing, especially the cut-and-paste and decontextualized quote dumping that often stands in for writing.

    5. Thanks to you both dmf and Kirsten for your comments. I think the idea of bringing in more case studies and fieldwork-based writing would help address part of the problem here. If we’re just asking students to find information and copy it into another context, that’s really just repetitive work. It’s no surprise that some end up using tools like chatGPT. Some of the suggestions for dealing with this problem include adding more specificity to assignment prompts…and linking to fieldwork-based assignments could be one way to do that. Although not all assignments can have that component.

  3. Just a quick update about the ‘hallucination’ and where it may have come from. I do think it’s related to what Ted Chiang writes about in the New Yorker piece. So there is someone named David Wilkie who has done some anthro and conservation biology-type of work. But not in Cabo Pulmo or Mexico. And one of the people that I interviewed for my doctoral work, and who appears in my dissertation, goes by the name “Wilkie.” I think chatGPT basically filled in the blanks between those two points and that’s where the fabrication came from. Although I still find it amazing how confidently it spit out those fabricated details…