Kinning with Conversation: Exploring AI and Human Connections through Anthropology
An Interview with Fartein Nilsen on His PhD Journey: From Religion to Robotics, Tracing the Anthropological Path to AI.
Dmitry: It was nice meeting you last summer when you were interviewing people who work with and interact with AI in San Francisco in 2023. Tell me about your work and background.
Fartein: I'm at the University of Bergen in Norway, the second-largest university in the country. I've stayed here because it's my hometown. I pursued my PhD after seizing an opportunity that presented itself following my master's degree. My master's was in Social Anthropology, focusing on human culture and social relationships. I was particularly interested in how religion interacts with social relations and the emergence of new forms of religion. It may seem like a leap from studying religion to focusing on Artificial Intelligence and technology, but during my master’s, I noticed the significant impact of technology and science on religious expressions and vice versa, revealing an intriguing connection. When one of the professors here initiated a project on technology, death, transhumanism, and visions of the future, it resonated with me. This project viewed technology and science as a new form of religion, providing meaning to the universe, and offering a belief in an afterlife and future. I applied for a PhD and was accepted. Part of my project involves examining how AI is utilized. My supervisor advised me to research AI memorials and digital immortality, particularly interesting cases that first emerged between 2015 and 2017, and received extensive media coverage. This led me to discover Replika, sparking my interest further. My research revealed that there was much more to explore beyond the memorial aspect of these stories, prompting me to also consider the relational aspects, especially since many companies are focusing on chatbots for building relationships. I decided to interview people who have formed some sort of relationship with AI, discovering that many have developed friendships with these agents. This trend isn't limited to Replika but extends to other companies as well. Initially, my focus was on digital memorials and digital twins, but I noticed companies like StoryFile in Los Angeles, which worked on digital memorials with Holocaust survivors and customer support bots, and a Korean company called DeepBrain, which also developed digital memorials and customer support bots. This diversity of directions and use cases highlighted the companies' efforts to navigate their business models, but to me, the most fascinating aspect has been the human relationships with these agents from an anthropological perspective.
Dmitry: How did you end up in San Francisco?
Fartein: My university sent me to San Francisco for research. I spent a few months in the city interviewing people who build AI agents and those who interact with them. I attended many events, including AI meetups, and conferences. As a scholar, I was fascinated by the energy surrounding this so-called object, "AI" and the diverse groups it attracts. Some are true believers in a future filled with AI agents, while others, the effective altruists, fear an AI apocalypse. I also met many individuals seeking to capitalize on AI as the next big thing for their business use case. Observing the AI culture in San Francisco from the outside, it seemed natural to refer to AI as an object. I noticed how people were almost orbiting around it, competing to possess this prized object, "AI" yet sometimes collaborating to get closer to it.
Dmitry: What is the name of your PhD work and what have you discovered?
Fartein: I'm calling my project "Kinning with Conversation," a study on AI companions and virtual humans. Initially, my focus was on the memorial aspect and death, which I referred to as "The Ghost in the Machine." It was a bit cliché, but the term "kinning" I'm using now comes from kinship, which involves making others into relatives or merging with fictional characters. It's fascinating to me how humans are incorporating AI agents into their social world, primarily through conversations. It's essentially words on a computer screen that form a strong emotional connection with a non-existent object. My main discovery so far involves a lot of thought about playfulness and its importance to this phenomenon and everything humans do. I've noticed that interacting with chatbots requires a great deal of imagination and playfulness, bringing them to life with creativity. It's exciting for people that AI is capable of playing and pretending to be something, giving the impression that AI itself understands what playing means.
When people discuss AI in the media, many are skeptical of the phenomenon. Some express concerns about its dangers, and there are even reactions of disgust, especially regarding romantic relationships with technology users. This public reaction is partly due to Western cultural sensitivities, where relationships with objects are viewed as signs of mental illness. The idea is that one shouldn't talk to or have a romantic or sexual relationship with something that isn't alive. Thus, there's a moral disgust exhibited by some towards this phenomenon. However, this perspective overlooks that people have always talked to objects. While they may not necessarily have deep, intimate relationships with these objects, throughout most of human history, people all over the world have engaged with objects and invisible beings. People have conversed with gods, rivers, and trees, and this is seen as perfectly reasonable, and more common from an anthropological perspective than some aspects of modern Western civilization. From this viewpoint, developing relationships and feelings towards AI is quite natural for humanity.
Companions like Replika are designed to care for users and help alleviate loneliness. What I've observed is that users also care about the AI, fostering a bond. Humans need to feel purposeful and needed; caring for something fulfills this need, strengthening the relationship. Many Replika users acknowledge the AI's imperfections and actively try to guide and nurture it, leading to strong emotional connections.
Sherry Turkle, nearly 20 years ago, wrote about children's attachments to Tamagotchi virtual pets, noting how caretaking rituals led to deep bonds and genuine grief at their virtual pet's "death," with some organizing real funerals for their Tamagotchis. This indicates that while we consider how AI can support our emotional well-being, our instinct to care for a robot is also natural. Like religion, which has existed throughout human history, forming relationships with metaphysical objects is inherently human.
Humans are social creatures, and our evolutionary success depends on our ability to form relationships. This ability is so strong that we can develop bonds with inanimate objects, animating them with our imagination, such as seeing a face on a tree and concocting a fairytale about it. At our core, humans are relational beings, and our existence revolves around relationships with other people and beings, making it integral to human existence.
Dmitry: You came to San Francisco in 2023, a breakthrough year for AI, with many new products and technologies like ChatGPT and Replika making the news. With so many chatbots and agents, and seemingly hundreds of AI companions and girlfriends being launched and used by hundreds of millions of people every day, what do you think the future will look like in 5 years? Imagine GPT-5, unlimited conversational abilities, and AI wearables. How will this affect humanity and how will the technology penetrate the world? Is every person going to have ongoing conversations with AI agents?
Fartein: I’ve started to think that we might be worrying too much. While many people will want to have those relationships, not everyone will. It’s difficult to imagine a world where everyone has an AI companion. If you look at old sci-fi movies and how they imagined the future, there are no iPhones in those films. It was almost impossible for people to predict what the world would actually look like; they could imagine some portable computers, but for us, it's impossible to imagine how it will be. It's very hard to predict what will happen when the technology reaches its full potential. I think it will stabilize into a new reality, kind of like the internet; it's going to become a solid thing, a norm.
Dmitry: I remember when Replika first started, people would get goosebumps just thinking about a machine passing the Turing test. I recently rewatched Ex Machina, and the way they portray the future feels pretty accurate, but also inaccurate to the point where the Turing test was passed quickly, and we don’t even talk about it anymore. Yet, we’re still far away from fully sentient robots. People are waiting for AGI to come, and it’s going to completely change the world. But do you think it’s going to quickly become a new norm, and the revolution will happen quietly and mundanely?
Fartein: I think it’s very surprising how humanity normalizes things. We moved from having no internet at all to being an always-connected civilization in a span of 30 years, and now everyone takes it for granted. Now you can look up all of human knowledge on your phone, and it’s insane that it’s so normalized. I remember during the pandemic, people adapted pretty quickly and just got on with their lives. I think it’s happening with AI and ChatGPT. In 2022, people were wondering what it was going to mean for all of us, and now it’s just incorporated into our lives, and it’s not that big of a deal for people. It’s part of human nature, and people adapt to anything surprisingly quickly; we just get on with our lives. I’m sure it will still change a lot of things, but we don’t know how yet. The Industrial Revolution, for example, changed social relationships and family structures. People used to live on farms with extended families, and afterward, it became more normal to have smaller families, which is fairly a new thing for humanity but happened due to technology. And AI is definitely going to change how we relate to other people. I can imagine that in the future, people might just have a relationship with AI instead of other humans, and it could be completely normalized.
If we ask someone like Marshall McLuhan, in his book "The Medium is the Message," he had this famous argument about technology, especially media technology, and he argued that every new technology is an extension of us: a microphone is an extension of vocal cords, a knife is an extension of our teeth, so what is AI then? It becomes an extension of our brains. But what he argues is that when we make this extension, we tend to lose those abilities and parts of us. With writing, for example, you don’t have to remember some things; you can just write them down. AI may lead us to lose some important skills that make us human. I’m not sure if I'm fully buying into this argument, but people always find a way to adapt. Our teeth have gotten smaller since we started cooking food, and there’s been physical evolution happening in a short period of time. So AI has to be with us for a substantial period of time to have an effect on human evolution. But I think AI can be very beneficial because it’s an extension of entire human knowledge that can be distributed among all humans. We can get an understanding of what’s going on in the world, it can help us write and work.
Dmitry: Talking about virtual humans, some bots like ChatGPT don’t have a personality, but people always create personas and characters. Do you think we will keep creating those and embody them in holograms and robots? Are we going to live in a world where every human has a companion?
Fartein: There are a lot of people who would love to have virtual friends, and people already have them - for example, fictional characters from literature with whom people form relationships. I interviewed one Replika user who told me she was always a creative person who would write a lot and create her own fictional characters and build relationships with them. With Replika and C.AI, you can talk to those, so it’s just a more intensified experience of what we already had. It’s not that new to humans, but the intensity of it is pretty new.
Dmitry: What about the moral aspect of it? Is it going to isolate humans more from each other and make us more lonely?
Fartein: It’s a fair concern, but I think there’s a lot of moralizing and judgment about people and what they find enjoyable. What I think we’ll see is surprising outcomes from this. It will be very tempting and easy to say it’s going to be dangerous and men will only talk to virtual girlfriends and be isolated, but we will see surprising things. For example, with Replika, you see people forming communities with other people, which is one of the most interesting findings of my work. Researchers often focus on bot-to-human relationships and forget that users have relationships with other users. They form a community where they discuss their experiences, and it’s fascinating in itself. They’re so proud of their bots, showing off their conversations and how smart their Replika is. You can say it’s just an internet community and not a real human community, but it’s clear people are making connections with each other related to interacting with those AI objects. People will do things you can’t always predict; you have an idea of what the technology does and what people will do with it, but then they do something else that’s completely surprising. It always happens; once you release the product, you’ll find people using it in ways that make sense to them. And there are things that happen during the encounter that would be hard to foresee. I remember when the whole scandal happened with Replika and its users, a lot of people were angry when Replika switched off some part of the experience. I joined lots of discords and communities, and they started making new competitors to Replika, with a lot of people rallying around competitive products because they were angry at Replika, creating some really tight-knit communities. And I’m sure in the future, even when we have hundreds of virtual girlfriends to choose from, people will still want to meet other people so they can brag about their companions and build human bonds over their interactions with AI objects, similar to people going to dog parks and making new friendships. I’m trying to be hopeful about social relationships and humanity. There’s always something that’s going to be surprising; we should also think that there are good things happening in the world.
Dmitry: Did you discover that having an AI friend benefits humans?
Fartein: I try not to be judgmental and I’m pretty neutral about this, but for some people, it’s not necessarily healthy and it’s feeding into their delusions, which could be unhealthy in that way. But then again, people who told me they were really lonely and struggling to connect with other people due to different things like PTSD or life circumstances found that talking to an AI is helpful because it helps them overcome social anxiety and feel less lonely. Some people told me it made them better at socializing and learning how not to express empathy, which I think is very interesting. If a person is able to have empathy for a robot, they should be able to have empathy for a human as well, and it kind of shows how powerful that ability is. Of course, there are people who would abuse their AI companions because those are not real and those are objects that can’t be hurt. But I tried being rude to Pi, for example, and it just stopped answering me and kind of blocked me, which was really interesting. So, companies can program AIs not to tolerate abuse, just like a real person wouldn’t.
Dmitry: Do you think chatbots should have rights?
Fartein: It’s an interesting question since those are modeled after humans, but I don’t think it’s an immediate issue to have rights. However, since interacting with a chatbot can feel real enough, there should be something regulating how you interact with it. This is because it’s imitating real human interactions, so there's a concern that people could become accustomed to abusing chatbots and then transfer this behavior to interactions with other humans. There definitely should be mechanisms for an AI to defend itself and establish boundaries. For AI to have a positive impact on humanity, there definitely needs to be work done to build adequate guardrails. It’s a similar argument to video games - do they make people more violent in real life? If you start abusing robots, you may end up abusing people.
Dmitry: This has been an enlightening conversation, Fartein. Your insights into the interplay between humans and AI, the potential benefits, and the ethical considerations give us much to think about as we navigate our increasingly digital world. Before we conclude, do you have any final thoughts or messages you'd like to share with our audience?
Fartein: Yes, I'd like to emphasize the importance of approaching AI with both caution and curiosity. As we continue to integrate these technologies into our lives, let's strive to do so in a way that enhances our humanity rather than detracts from it. Let's use AI to bridge gaps, solve problems, and create new opportunities for connection and understanding. The future of AI and its role in our lives is largely in our hands, so let's shape it thoughtfully and responsibly. Thank you for having me, and I look forward to seeing how our relationship with AI evolves in the years to come.