I’ve often been asked why an anthropologist studies digital technology. It is perhaps easier to answer that question today, when AI endeavours to teach, assist, entertain, drive and even heal us—essentially, to be like us. In order for AI to achieve these goals, it must learn who we, as humans, are—our actions, preferences, needs, thoughts, knowledge and values
In the late 1990s, I encountered Nico in a chat room— at a time when “age/sex/location” was the standard icebreaker. As we shared our locations, a surprise awaited. “I didn’t expect that,” remarked Nico from Croatia upon learning that I was from Serbia. We switched from English to “our language” and an online romance began. With it, my professional path of digital anthropology also began.
Falling in love online before the release of the movie “You’ve Got Mail” in 1998 was so unconventional that I needed to understand what was happening. I thus embarked on studying social relations online and soon authored what is considered one of — if not the very first — academic articles published in Serbia on the sociocultural aspects of digital technologies (Sociologija, 41/2, pp. 187– 200, 1999).
The internet has since served as a distinct mirror through which I reflect, observe and anthropologically analyse the reality around me. My exploration of the co-construction of technology and culture led me to research and teach in the U.S. and Europe, delving into topics such as internet use during wartime (University of Belgrade), digital research tools (Royal Netherlands Academy of Arts and Sciences), digital humanities (University of Oxford; University of Amsterdam), digital scholarly workflows (Pennsylvania State University), virtual embodiment in the metaverse (University of Minnesota), digital research methods (Illinois Institute of Technology), and the sociocultural aspects of artificial intelligence in applied research (Silicon Valley).
Viewing modern AI research through an anthropological lens is crucial, as AI research often oversimplifies human behaviour into rudimentary models, overlooking its true complexity
I’ve often been asked why an anthropologist studies digital technology. It is perhaps easier to answer that question today, when AI endeavours to teach, assist, entertain, drive and even heal us—essentially, to be like us. In order for AI to achieve these goals, it must learn who we, as humans, are—our actions, preferences, needs, thoughts, knowledge and values. The current development of AI is thus like a gigantic ethnographic project aimed at learning as much as possible about humanity.
This global AI ethnographer isn’t an anthropologist and often isn’t even human, but it bears similarities with anthropology in its early days. Just like AI-focused computer science today, anthropology was once arrogant in assuming that it could fully capture human worlds and worldviews, objectively understand and interpret them, and guide the “less-advanced” along the presumed, unquestioned evolutionary line of progress.
AI-focused computer science thus still needs to learn what anthropology learned a while ago—that such a mission is not only impossible, but unethical and detrimental. It took decades for anthropology to evolve its current principles, such as reflexivity, community-based participatory research and cultural relativism.
Viewing modern AI research through an anthropological lens is therefore crucial, as AI research often oversimplifies human behaviour into rudimentary models, overlooking its true complexity. This oversimplification leads to extreme visions of a utopian or dystopian future, under the mistaken belief that these outcomes are inevitable. Ultimately, it is us humans who must actively shape our future, including the technological aspect, rather than passively accepting it as predetermined. Let us thus approach our co-construction with digital technologies, including AI, with the mindset of anthropologists—or as online lovers—embracing the possibility of surprise and acknowledging that everything could always be different.