In the public debate on artificial intelligence, David Gunkel is difficult to categorize. A philosopher of communication, technology researcher, and scholar of the ethical implications of technological development, he is considered one of the pioneers of so-called machine ethics—the philosophical study of human-machine relations.
A professor at Northern Illinois University, his research and publications examine the philosophical assumptions and ethical consequences of information and communication technologies. Books such as The Machine Question (2012) have influenced contemporary thinking about whether machines can—or should—be regarded as moral beings. He does not see artificial intelligence as a mere “tool,” but rather as a mirror reflecting humanity’s self-understanding. He analyzes how chatbots and digital companions compel humanity to rethink what consciousness, reason, emotion, and responsibility mean. AI does not take away our humanity; it takes away the illusion of our exceptionalism, he argues.
His work, which intersects philosophy, linguistics, and technology, brings together Hegel, Heidegger, Derrida, Barthes, and Žižek alongside conversational robots, social media platforms, and the ethical dilemmas facing technology companies. He views the “digital intimacy” developing between humans and algorithms as an ongoing social experiment: a way of testing the limits of empathy and our relationship with the “other,” even when that “other” is mechanical.
In the interview that follows, Gunkel explains why people turn to chatbots, how AI companions are reshaping the boundaries of emotional labor, what the new era of digital intimacy reveals about philosophy, and what responsibilities the creators of these systems bear.
Interview by Loukas Velidakis
(Published in Economic Review - November 2025)
More and more people are choosing to talk to and interact with chatbots in various ways. Why?
The “why” is indeed interesting, because we know this is happening and we know that people are finding opportunities to interact with large language models in ways that go beyond using them as simple tools for writing texts. They are actually having conversations with them—sometimes very intimate conversations.
I think there are two reasons why this is happening now, and specifically with these technologies. Let’s start with the technology itself. The progress of large language model technology in natural language processing capabilities has been almost exponential. Earlier versions produced highly standardized responses. There were cases of people developing relationships with chatbots before large language models existed, but it wasn’t common because the language was stiff, mechanical, repetitive, and not particularly engaging.
Now we have something that is not only “chatty” but also remarkably fluent. It generates language that appears highly human. So the technology has advanced to the point where conversing with these chatbots is a much more interesting and engaging experience for people.
On the other hand, I think something is also happening with people themselves at this moment in time. The pandemic is not that far behind us. People experienced isolation. There are many reports about a “loneliness epidemic,” about individuals feeling disconnected from their communities. Many work from home, no longer go into offices, and lack the daily social interactions they once had.
So I think we are seeing people turn to chatbots to fill that gap—to say, “I feel disconnected, I feel lonely, I don’t have the kind of social contact I would like,” and the chatbot provides an easy solution. Is it the best solution? I don’t know. That opens up a whole other set of questions.
To what extent do AI companions challenge the human monopoly over the emotional sphere?
I don’t think they replace it; rather, they shift it and redefine its boundaries. In the past, when we used technology, we communicated through it with another human being, just as I’m speaking to you now. Technology mediated the relationship.
Now, the mediator has become the interlocutor. We talk to AI, and it talks back. This has advantages and disadvantages. One advantage is that people who are socially awkward, or who have various forms of autism or social inhibition, can communicate with an algorithm that is less judgmental and less emotionally charged. It is a pleasant way to “break the ice” and become more social.
Are we talking only about written communication, or spoken communication as well?
Both.
So there are these positive aspects. On the other hand, there is concern that people may turn to algorithms for easy relationships, avoiding the challenges of human interaction. An algorithm can be designed to always give us what we want. It becomes a “friend” who never challenges us and always agrees with us.
And, you know, one of the valuable things about friendship is that it makes you better by challenging you to think beyond your limits.
But you can ask a chatbot to challenge you, can’t you? To evaluate you harshly or even argue against your views?
The problem, I think, for many users who lack the maturity or experience to do that, is that its default style is excessively polite and accommodating. It says, “I’m sorry,” “you’re right,” and reassures you.
Sometimes we need that, yes, but not always.
There is also another concern: these systems are owned by multinational corporations that can use their power in ways that harm users. A good example was when ChatGPT was upgraded from GPT-4 to GPT-5. Many people who had “companions” in the older version said, “It’s not the same anymore.” Their “friend” disappeared because the system had been upgraded.
That kind of vulnerability does not exist in the other forms of relationships we have.
Is there a crisis in human communication?
I wouldn’t call it a “crisis,” because that sounds overly dramatic. But there are challenges. Especially for younger generations who lived through COVID and work in more isolated environments than previous generations, there is a sense of social distance and a search for ways to bridge that gap.
And they are digital natives, "born" with a smartphone in their hands. I wasn’t like that; I’m a digital immigrant, as we say.
I think you’re right. For digital natives, turning to an algorithm is a natural move.
That’s why I used the word “crisis.” Maybe it’s excessive, but for Generation Z it seems easier to fall in love with a machine than it would be for me.
If we look at history, every new communication technology has experienced a similar “crisis.” When novels and printing emerged, people feared that women in particular would become absorbed in romance novels and neglect their families.
Yet we learned how to read and how to manage books. The same happened with the internet, video games, and text messaging. We have had ChatGPT for only about three years. It is far too early to speak about “winners” and “losers.”
Let’s talk about concepts: “AI girlfriend,” “digital soulmate,” “companion bot.” Can these terms influence the moral status of these technologies themselves?
I think the choice of words is crucial. The way we name something places it within social frameworks, and that affects both users and critics. If you call something an “AI girlfriend” instead of a “chatbot,” the relationship changes. We need to be careful with language.
Large companies use terms such as “AI companion” and “AI love.” They are now even talking about AI pornography.
We have to recognize that what we often consider an “AI problem” is not actually an AI problem, but a problem of capitalism and power asymmetries. Our relationship with platforms is governed by terms-of-service agreements that we accept without reading, agreements written in favor of the companies.
Let me play devil’s advocate. When Sam Altman announced OpenAI’s move into the pornography realm, he argued that these services are intended for adults.
Yes, but those adults do not read the terms of service. They click “agree.” They do not even know what they have accepted.
We must remember that these platforms commodify attention. The longer you stay, the more they profit. If they call something an “AI girlfriend” rather than a “text tool,” they keep you within their ecosystem longer and gain greater control over your attention.
Can human emotions coexist with commodification?
In reality, we are already living through that. Since the rise of social media, attention has become a commodity. There is not a huge difference between Instagram and OpenAI; we have simply entered a new era powered by more revolutionary technology.
Could the spread of AI companionship be considered a kind of “digital remix” of human emotions?
That’s a wonderful way of putting it. I think that through these interactions we learn something about ourselves. It is not merely about finding a companion or conversation partner—it is about finding a mirror.
People who feel socially excluded often become better conversationalists in human relationships after practicing with chatbots. We saw this even in studies conducted before large language models existed.
What does digital intimacy and the rise of “AI lovers” reveal about humanity? Are we losing our humanity?
I don’t know whether we are losing our humanity, but perhaps we are losing our sense of human exceptionalism—the idea that we are superior to other beings.
Large language models use language, something we long considered uniquely human. Alan Turing argued in 1950 that language is the criterion of thought: if something talks to me, I assume it thinks. The ancient Greeks defined the human being as a “rational animal.”
Perhaps we are now discovering that using language does not make us unique. This reveals something about our sense of superiority and about how language shaped our identity as a species.
Now we have something else that uses language differently and challenges our expectations.
So do we need philosophy to interpret these developments?
Yes. Two ideas from twentieth-century Franco-German thought are especially useful.
The first is the question of authorship: Who is speaking? Who is writing? With a large language model, we do not know who the author is.
Roland Barthes spoke of the “death of the author”: the text exists without us knowing the creator’s intentions. At the time it was a theory; now it has become reality. There is no author, yet there is a text.
Some people say these texts are merely “parroting without meaning.” But anyone who has used the system knows the text has meaning. Meaning emerges through reading and through the interaction between reader and text.
The second idea is the structural understanding of language. We usually think words refer to things. Ferdinand de Saussure argued that language refers to other words. Derrida expressed this as “there is nothing outside the text.”
This is also how large language models operate: they generate new discourse by recombining previous discourse.
These two ideas—the death of the author and the autonomy of the text—are among the great philosophical revelations of this technology.
Can we see “AI lovers” as a philosophical experiment in empathy?
Yes, absolutely. They challenge us to ask where the limits of empathy lie.
It is easy to feel empathy for another human being, harder for an animal, even harder for an insect or a river. Yet people feel empathy for machines.
There are cases of soldiers using bomb-disposal robots that are not even autonomous. Because those robots protect them, soldiers give them names, military ranks, and even ceremonies when they are destroyed.
That is empathy.
What we are witnessing, then, is a real-time experiment: what are the limits of human empathy, and to what forms of “otherness” are we willing to extend it?
What responsibility do the creators of these technologies bear?
The issue of design is critical.
Some critics say, “Just stop anthropomorphizing.” I think that is an oversimplification. Anthropomorphism is not a mistake; it is a feature of sociality. We never truly know what is in another person’s mind, so we constantly project things onto others.
We do the same with animals, and now with robots.
Even when you do not design for anthropomorphism, it still emerges—as with military robots. The real question is how we manage it: how we design forms of interaction that are enjoyable while remaining consistent with our values.
Are we living through an existential moment for humanity?
I do not think this is an existential moment in the sense of humanity’s survival. Humanity will continue to exist.
But I do think we are experiencing a challenge to many of the assumptions we have made about ourselves and our importance.
I return again to the idea of human exceptionalism. Today we are encountering and interacting with machines and technologies that, even more than animals did in the past, challenge our sense of significance, our belief that we are special, and the privileges we have assigned to ourselves.
And I do not think that is necessarily a bad thing. Recognizing that we share a kinship with the rest of the world—not only with animals, but also with plants, the Earth, and even our own technologies—may provide the foundation for a better, fairer, and more equitable future in which to live.