Interview: David Gunkel on AGI Ethics and ‘the relational turn’ – Part Two

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 “entities” 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 second chapter, Gunkel and Miller dive deeper into the relational turn and discuss influential figures in the field.

KB: It sounds as if [the relational turn], to some extent, is behaviorism by another name. And I wonder how it differs, really, from a basic behavioralist approach.

DG: Yes, that’s a really good question. In fact, one of the individuals who’s written in this area, and is in alliance with the work that I’ve done on the relational turn, is John Danaher from the University of Galway. He created a framework called Ethical Behavioralism which is precisely this. And it has certain affinities with the relational turn. 

But behavioralism, inevitably, is based on having certain properties that exhibit behaviors. And the relational turn, if we take it seriously, is not grounded in the existence of pre-established behaviors, but rather it evolves out of the social interactions, oftentimes, prior to being able to assign behaviors to the other. It recognizes that the behaviors we assign to others are actually a product of the relationship…

So, I think the affinities are there: the focus on extrinsic interactions and activities is driving the process of decision-making. But my worry with behavioralism is that it still is grounded in psychological properties.

KB: That would seem to open the door to entities having ethical standing. I think [Silvia] Benso is a big fan of [all things] having ethical standing … whether they … have behavior or not — the pebble in the path has ethical standing by virtue of where it came from, and how the Earth was changed by virtue of it getting there, and how the lives of plants and animals and humans around it were changed by virtue of it being there. 

It’s not an [anthropocentric] ethics … becomes more than an anthropocentric domain if all things have ethical standing … Do you subscribe to that or is that something that you’re trying to get away from in a certain sense? Because it’s kind of more essence-based?

DG: No, I think I’m very much in alignment with that. I will say that, for me, reading Benso’s book, The Face of Things, was really a crucial experience, because it helped open up this opportunity to think in this fashion, and to really push [Emmanuel] Levinas much further than he would be willing to go. As you know, in Levinas’s own works, he is very anthropocentric. 

And [Jacques] Derrida pushed him on this in “The Animal That I Therefore Am.” Matthew Carco [IERCD] pushes him on this, in the work that he does with environmental ethics. And Benso pushes him on the artifact. And Levinas needs to be pushed, because there’s no reason within his own system that allows for that anthropocentrism to be grounded and justified. 

The idea that otherness and alterity are nouns is the source of the problem. These are not nouns — they’re not substances — but they’re verbs. Alterity is about the effect of alterity; it’s about the event of alterity, as Derrida might call it. Thus it is the case that anything can supervene in such a way that it is other, that it interrupts your—what Levinas calls—”ipsity,” your sense of self-enclosure. 

So the pebble in the road can do that, the animal can do that, the plant can do that, the robot can do that.

And I think if you radicalize Levinas far enough in that direction, following Benso and others, you do have a situation where you have a radically decentered way of doing ethics, that looks at ethics as something that is no longer centered on the human, or even on human capabilities, whatever we want to call it. In fact, what I like to say about Levinas’s own work is that it’s not a moral centrism at all. It’s a moral eccentricity, where you have many centers, and not just one. 

And I think if you look at non-Western traditions, especially indigenous traditions, where you have this notion of kinship with the things around you, so that you share kinship not only with other humans in your community, but you share kinship with the animals, with the plants in the environment, with the dirt, with the mountain, with the waters, all these things then start to matter in ways that create a sort of moral stance towards the entire universe…

KB: I probably, like a good therapist, shouldn’t be telling you about my background coming to this, but I read Benso about 15 years ago. And she changed my life. I mean, to hear you talk so, in such a lively and enthusiastic way, about not just Benso, but the marriage of Levinas and Heidegger.

That what she did is really terrific. I’ve been working with Sheldon Solomon from Skidmore on these issues, almost since that time. But only recently in a creative way on [AGI Ethics News, but] we were writing about it in 2017. And I think I’m not going to take up much time here, but I just wanted to say that the fact that you acknowledge Benso’s what I call elegant gymnastics at bringing Heidegger — the best of Heidegger — and Levinas together, I thought is really great. And the philosophical community has not stood up for that, apart from you, that I’ve seen … since the book was written in 2000.

DG: Yes. Can I just say that when I have these conversations — and this is, for me, very interesting and revealing — that I’m the one who has to introduce the interviewer to Benso. I will mention her work, and people, unfortunately, don’t know it. And I have to sort of do the background work to inform people of what she was doing and why it was so important.

So it’s incredibly gratifying for me to have you mention Benso even before I had gone in that direction, because it’s nice to talk to someone who also appreciates that intervention that she created back in the 90s. I think it’s a remarkable book, and I wish it was read more widely.

KB: I think it’s essential reading, particularly in the West, where it takes a kind of analytical approach, even though she’s a continental it should be essential reading. It brings Eastern tradition and indigenous tradition into a perspective that, for… for me personally becomes pragmatism.

DG: Oh, yes, that makes a lot of sense.

KB: And so, yes. All right… So there’s a growing camp that says that we should be looking at … this is back to the pragmatics of AI and AGI in particular … the idea that we should be looking at harm-doing rather than at virtue-building.

You probably are aware of Shannon Vallor’s work in Technology and the Virtues which I found another really good [ethical framework]… what I took to be … the first try at techno-virtue ethics. And in fact, we use that in our creative work. 

How do you stand on the issue of do no harm versus be good? Is there a way that we can build AGI with one or the other, or how do you see that?

DG: Yes and no. It’s a really important question, and I also find Shannon’s work really useful and innovative, because I don’t think anyone up until that point was really seriously thinking about a techno-virtue ethics, and she’s given us one, or at least the beginning of one, which I think is very important for the conversations in AI ethics.

You’ll notice, too, that a lot of the AI ethics literature and discussions that are going around, not just in academia, but also in the industry, is really focused on reducing harm, or controlling for harm. And that, I think, sounds good, because no one wants to be harmed, and no one wants to engage in creating technologies that produce harm.

But it doesn’t focus on doing good, necessarily. It focuses on just restricting bad outcomes. And I think that is a very limited way of looking at the opportunities and the challenges that confront us in creating these technologies and deploying them in society.

And so, I think a virtue-focused ethic might be one way to look at this differently. And ask, not, how do we minimize harm, but how do we engineer systems and implementations that cultivate good, that cultivate justice, that cultivate value? And that question, I don’t think, is being asked enough, because the harm stuff gets a lot of the attention, and this seems to be kind of a marginal note. Or even a footnote in some cases.

So, I would be interested to see how, especially [for] the field of AI ethics, what new possibilities would there be for doing the work in AI ethics? If the focus moved away from safety and harm, which are negative kinds of attention, more in the direction of the positive: How do we engineer for good? How do we engineer for success? And that may shift the emphasis and the way that we approach a lot of these dilemmas and problems.

KB: The objection to that that I hear is that it can’t be built.

DG: Right.

KB: Constraints can be imposed. Guardrails can be written. But engineers have a hard time conceiving of what good is. [So do we all!] So there’s this tension between the empiricists, who are actually trying to create usable technology, and us, people talking about the ground of being… Do you have thoughts on that? I mean, is there a way around that?

DG: Yes. So, I don’t know that there’s a way around it, but I know there are better and worse strategies by way of engaging with it… and creating operative frameworks for it.

I think one of the ways that we’ve typically tried to resolve a lot of these dilemmas is to put the onus on the engineers, and say, well, “we just gotta design these things better. We just gotta engineer these things for safety. We have to engineer these things to reduce harm.” And, we put a lot of responsibility on the shoulders of developers and designers, and say, you need a better design ethic. 

So it’s all design side. And that’s where the focus on ethics is. 

I think that’s maybe half of what’s going on here. There’s a lot that happens on the user side. Once these things get out into the world, what users do in response to these new affordances and these new capabilities oftentimes reshape them. There’s a very good line, in one of the cyberpunk novels of William Gibson, where he says “the street finds its own use for things, things the manufacturers never anticipated.” [“Burning Chrome” 1982] And this is true of a lot of the technologies we’re dealing with right now. 

We invent various tools, we put them out into the world, and we find out that users don’t want to use them the way designers anticipated, and that we do something different in the process. 

And so I’m very interested in user ethics, and especially a user virtue ethic. Like, what is it to be a good user? And what is it to engage in using these technologies that make available to us better or worse choices, with regard to outcomes. And I don’t know that there’s been enough focus on the user side because a lot of the work in AI ethics is really design-focused. And for good reasons. I understand why. But I do think that we need to balance that with a similar attention to what happens on the user side.

AGI Ethics News

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.