In an exclusive interview for AGI Ethics News, KB Miller sat down with philosopher David Gunkel to discuss the ethical implications of treating artificial general intelligence systems as “things” and to explore a radical shift in how we approach the moral standing of artificial intelligence: the “relational turn.”
David J. Gunkel (PhD – DePaul University) is an award-winning author, educator and researcher, specializing in the philosophy of technology with a focus on the moral and legal challenges of artificial intelligence and robots. He is a Professor of Media Studies and Department Chair at Northern Illinois University. He’s written numerous books including Person, Thing, Robot and Robot Rights.
Due to the length of the interview, we’re splitting it into four parts. In this first chapter, Gunkel explains the relational turn and argues that moral status emerges extrinsically from our ongoing social interactions with the entities around us.
KBM: It would be good to start with “the relational turn” and what it is, and its core claim, and perhaps how it’s commonly misunderstood.
DG: So, this whole business concerning what I and others have called “the relational turn” is really about how we decide who is worthy of consideration and respect, as a subject of moral concern, and what remains a mere object — a thing that we can utilize instrumentally as we decide and see fit.
And ethics really is about making this distinction. We decide. We look at the world around us, and we cut it up into little categories. These things are like us, they matter to us, they’re sentient, they’re conscious, whatever, and therefore they have moral status.
These other things — like the toasters, the cars, and the other tools in my world — those are just things. And they’re pieces of property. And the way we normally make this decision is, we look for a property or a set of properties that we believe are the ground [for] moral consideration.
And in the history of philosophy, this has been a moving target. If you go all the way back to ancient times, what was the criteria for moral consideration? Well, it was that you were the male head of the household. And that meant that your children, your wife, your animals, they were all your property.
And you could dispose of them as you see fit. Now, over time, we realized that was a rather spurious property — male head of the household — and we came up with other ones, like rationality under Kant, like consciousness more recently, or even sentience, in the work of animal rights advocates like Peter Singer and people like that.
“And if we are honest, we don’t know for a fact that other people are conscious.”
Now, this approach works pretty well. We’ve been doing this for a good 2,000 years or so, and it’s pretty much definitive of the Western way of thinking about making these distinction, these categorical divisions in our world.
The problem is that they’ve really run into trouble, if we take it rather seriously and rigorously. And you run into troubles in three domains.
One, the area of determination. How do you determine which property or exact properties are the qualifying criteria? And the fact of the matter is that this has been a moving target and we’ve not been really good about actually determining what we think is the qualifying property. It has, therefore, been open to some negotiation that has led to some disagreements about what qualifies something to be “someone.”
There are also problems with definition. So let’s just take the most recent property that people think is going to be the defining condition for AI having moral standing: consciousness. This is the situation with the AI welfare movement, which has recently gotten traction with Robert Long and Jeff Sebo and others. They’re all taking this precautionary approach, saying, well, we don’t know if AI’s conscious, but we should bet on it maybe becoming conscious, and therefore we have to accord to it certain rights and privileges that we would to other moral subjects.
The difficulty is that we don’t have one definition of consciousness. In fact, if you look at the competing theories of consciousness, we are at about 50 of them right now. No one has quite figured out what consciousness really is, and yet this is supposed to be the defining characteristic that lets us figure out whether something is morally relevant or not.
Lastly, there are problems with detection. All of these properties are internal mental states, or emotional capabilities that are inside the psychology of an individual.
And if we are honest, we don’t know for a fact that other people are conscious. This is Descartes’ problem in The Meditations. He said cogito ergo sum, I know I’m a thinking thing, but [then he] looks out the window and says, “how are these things thinking things?” “How do I know?”
And we normally go on some sort of behavioral signs or symptoms that we take as indicators of intelligence or consciousness or sentience being present.
But inevitably, this is the problem of other minds problem. And even if this is not this kind of insurmountable philosophical difficulty that is often advertised, it is enough to cede reasonable doubt as to whether the property exists or not in an entity, and therefore there can be considerable disagreement about who or what possesses the property or doesn’t possess the property.
So, the relational turn just says, okay, let’s just be more honest, then, about the way that we make moral decisions. This “properties” thing sounds good, but it doesn’t really provide us the ground that we think it does. And the fact that it has been rather changeable and a moving target over time, should tell us that, really, when we make moral decisions, we’re doing something a little bit different.
And what we’re doing is, in the face of others, whether it be another human, a non-human animal, the things of the environment, or an intelligent artifact, we are in relationship with it, and we build our decision-making about their capabilities through the interaction of that relation. And then, afterwards, we project the qualities or the properties into it that we think motivated this decision-making in the first place.
So, the relational turn just says, what if we were honest about the way that these decisions are made, and we recognize that the moral status of others is something that evolves not from an individual’s properties, but from the communal interactions that we have with others in shared reality and as a result, it’s not necessarily coming from an intrinsic state, but it’s coming from an extrinsic interaction. We find ourselves engaging with all these other things around us.
KBM: So those interactions like [Silvia] Benso [RIT] might refer to them as a kind of connectedness to the universe. [Martin] Heidegger might say it’s the Fourfold reality. This vast, infinite connectedness to the universe. Is that the kind of connectedness that you’re referring to when you say that, or is it something else?
DG: Yes, it’s interesting. It shares a lot of affinities with Heidegger, especially his later work on The Thing. It shares affinities with Karen Barad [UC Santa Cruz, Science & Justice Research Center], and the idea that relations precede the relata.
It also, has a lot of connections to Benso’s The Face of Things. And in fact, that’s really where, I think, for me, this resonates very strongly, because this is a very Levinasian way of approaching this—to look at the way in which the other, as something that interrupts your selfhood. It exposes you to the outside, to the exterior. That is the kind of minimal, baseline experience of otherness that causes us to need to respond.
And respond … responsibly.
“The relational turn doesn’t give us a different morality, it just says, here’s another way of looking at it.”
And I think that’s really what the relational turn is trying to get a handle on, is that, even before we know cogito ergo sum, I think therefore I am, or any of this stuff, we are already alongside others who we need to make responses to — we need to indicate in our response to them, how they stand in relationship to us.
And that is a precognitive decision where ethics is first. And not second.
KBM: Let’s go back for a minute to the idea that relational turn is misunderstood. Have you found elements of it that are misunderstood … in print.
DG: Yes, there are a couple of ways that it’s misunderstood … let’s talk about three of them in particular.
So, one is that people think this is a simple binary—that you’re going to be forced to choose properties versus relations. But it’s not a binary. They are two sides of the same coin, they’re two ways of looking at the same ethical transaction. One proceeds from individual properties to ethics, the other proceeds from the communal interactions with others, where ethics precedes the individual, and that the individual arises out of that communal experience in which we already find ourselves alongside others.
So instead of this being a binary either-or, It’s more of a way of saying that for two millennia we’ve used properties as a way of explaining how morality works.
The relational turn doesn’t give us a different morality, it just says, here’s another way of looking at it that is maybe more honest about the phenomenological requirements and the epistemological restrictions regarding how we mobilized our moral decision-making and providing for us a more agile, but also a more responsive, way of responding to these questions about moral decision-making.
“And when you start to look at morality and moral philosophy, from a much more global perspective, you find that different cultures in different times and places have divided up the world into categories very differently from the way that we have done so in Western cosmologies.“
Another misunderstanding is the way that [the relational turn] is oftentimes discounted as a kind of relativism. Thus asking whether this expose us to a kind of “anything goes” outcome? That any relationship counts, and therefore, everything is possible, and therefore there’s no moral grounding.
Again, I think this is a misunderstanding of relativism. I think relationalism is not relativism, but it’s about the relativity in which we find ourselves, and that the moral conditions in which we operate are always asking us to make decisions, and those decisions often come with instabilities and uncertainties where we have to make a choice.
And those choices are not choices of absolute certainty, but they’re relative to what we know and what we can know, based in the circumstances that we’re engaged with.
So, one way to express this is that it’s not relativism, but it’s an ethical pluralism that recognizes that even when there are shared commitments and shared values, there are diverse ways in which people have resolved these questions.
And when you start to look at morality and moral philosophy, from a much more global perspective, you find that different cultures in different times and places have divided up the world into categories very differently from the way that we have done so in Western cosmologies.
And that we might learn something about our own sense of privilege and our own sense of making sense of the world, by looking at how other cultures—indigenous cultures, non-Western cultures, African traditions—have also negotiated these things, but have done so in ways that are remarkably different, and I think that can help us develop better understanding.
Now, some of the things that we thought were unassailable, some of the things that we thought were universal, turn out to be just as culturally specific as we think others are. And that kind of reflection on our own sense of privilege is actually very important here.
The last criticism or opposition people have given to the relational turn is that they’re worried that it’s going to be dehumanizing—that because you are basing [moral decisions] on how things are as opposed to what things are you could easily use the relational turn to deface or dehumanize other human beings, and therefore treat them as things and not persons.
This risk exists for any ethical system. Even when you work with properties, we can strip away properties very easily from different members of society and then dehumanize them in the process. So it’s not unique to the relational turn, this charge, or this risk of dehumanization.
But again, I think the relational turn explains how this is possible. It provides a more honest assessment of the ways in which power actually functions in ethics, that behind a lot of ethical decision-making, there’s politics. And that politics always involves exclusion and inclusion, and who gets to make those determinations — who actually gets to decide who counts as morally significant and what does not.
So I don’t think we escape [the problem of power] by trying to find some transcendental vantage point that is universal for all mankind, or humankind, in all situations and circumstances. I think we have to recognize that these boundaries are flexible, and they’re permeable, and that we negotiate these things again and again. But that’s ethics, right? And that’s the interminable struggle of ethics.
Continue to Part Two
In Part Two of our four-part series, David Gunkel explores indigenous kinship ethics, moral standing, and why AI ethics should move beyond harm reduction toward engineering for good.
“If you look at non-Western traditions, especially indigenous traditions, where you have this notion of kinship with the things around you… all these things then start to matter in ways that create a sort of moral stance towards the entire universe.”
AGI Ethics News covers the ethical and societal implications of artificial general intelligence (AGI). We bring expert insights and thoughtful analysis to help readers understand the challenges and responsibilities of emerging AI technologies.