Interested by this? Share it on social media. Join our monthly e–newsletter to keep up to date with our latest research and events. And check out our Friends Programme to find out how you can help our work.
We had a good event this week at Theos on Artificial Intelligence. It was wide–ranging and open–ended, largely because the field is developing at a remarkable speed and throwing up ethical and existential questions faster than we can answer them. John Wyatt and Beth Singler did an excellent job in outlining the contours of the debate and suggesting a few signposts.
This short essay (or long blog) is about one of those contours and one of those signposts.
The question arose in our discussion about whether it was in principle impossible for ‘robots’ to become persons, meaning, in the context, beings that we would be obliged to treat as if they were other humans. John Wyatt echoed my own view that until recently he thought that particular leap could never be made – an argument once made at length by Roger Penrose in his book The Emperor’s New Mind, with specific reference to that quintessential human quality, consciousness – but that now he wasn’t so sure. Perhaps machines could become persons. Perhaps one day they would. The question is: how would we know.
Park that thought and now read this.
If you are like me, you will find the prospect of such dolls nauseating, and I’m guessing that women will find it doubly so. The interesting thing is to ask why. After all, as Mary Midgley once argued, we should pay attention to (what she called) The Yuck Factor in our responses, as such an instinctive, emotional recoiling might tell us something profound about our more reasoned and legitimate ethical concerns.
I think my (our?) nausea at such ‘Real Dolls’ is because they demean women. More precisely, these ‘Real Dolls’ look, and presumably behave, like a teenage boy’s fantasy. They are expressions of male desire (and a pretty narrow example of male desire at that) rather than of anything intrinsically female. (“We are running a business and most of our clients have a certain wish list. The unfortunate reality is that that is rather idealistic”. Hmmm)
Sex with such a creature would, I image, be little more than elaborate masturbation. ‘Real Doll’ PR might talk about the “presence” and “personality” of their creations, with that personality chosen from a list of options. It claims that “AI gives people the tools to create that personality.” Nonetheless, the look and feel of such creations suggests strongly that what AI is in fact doing here is giving people a tool to project aspects of their own personality and desire, essentially to project themselves, not creating something different, new, or ‘Other’. Indeed, perhaps that is the very definition of Creation: the generation of something that is ontologically Other to you, something begotten not made.
This, as an aside, may also lie at the heart of the Yuck factor many of us feel with the idea of genetically selecting the characteristics of your offspring, another brave new prospect held out to us by technology. It begins attractively, indeed as an obvious good. Who wouldn’t select out genetic dispositions to serious and painful illness in our children? But then you move towards the thicker end of the wedge. Would you select out on the basis of intelligence, appearance, character? Do you not want a child who is clever, sporty, generous, loving? The problem here is that the offspring becomes not a person in him or herself but a project of the adult’s desires, loved not for being themselves, for their difference, their unique Otherness, but for their Sameness, their fulfilment of existing conceptions of what is good and valuable.
This moves us slightly closer to an answer to our presenting question. When might a robot be classed as a person? Perhaps it is when it is Other, “Other” in this sense meaning not simply a projection of existing human characteristics and sentiments but possessing those itself.
The wrong turn we often make at this point is to try and pinpoint a single, defining characteristic that marks us out as humans. If a robot’s got this, it’s made the grade. It is now a he or she.
I would contend that we should, instead, adopt philosopher Ludwig Wittgenstein’s idea of family resemblances to help us at this juncture. This is the idea that things are related to one another by a series of common features, no one of which is necessarily held by all parties. Much as members of a family may share “build, features, colour of eyes, gait, temperament, etc. etc. [which] overlap and criss–cross in the same way”, or games (his chosen example in Philosophical Investigations) may share similar family traits, despite being very different (e.g. card games, board games, ball games, etc.), so perhaps being a human person is marked out by certain family resemblances? Here are a few suggestions, in no particular order, of what they might be (none of them, note, physical – bipedal, upright, mammalian, etc. – but all based on us having physical presence):
A sense of finiteness and mortality, an awareness that once you didn’t exist and that one day you will die, a sense of personal narrative, a sense of inhabiting a moral landscape which makes demands upon you, an inclination to reproduce, a capacity and inclination to articulate your inner life… and one other characteristic mentioned below.
Even as I write this list, I can see some problems with it, but as this is a blog not a thesis I’ll let them stand for the moment. My point, in any case, is less to identify definitively those ‘family resemblances’ of human personhood so much as to suggest at that point at which a robot starts exhibiting some of these, we should consider according them the status of personhood.
How will we know? We won’t for sure, for precisely the reason that we are being asked to identify something that is intrinsically beyond and inaccessible to us, the very otherness of another being. But as with previous instances in (Western) history, when we have found ourselves evaluating whether other races have personhood (American Indians in the 16th century, black slaves in the 18th century) we would do better to err on the side of generosity.
A concluding remark: thinking along these lines about when is a robot a person should, self–evidently, help us think about what makes a person a person. The tentative list above is part of that, but the very act of thinking – being concerned – about the Other is also part of it. Why is a ‘Real Doll’ so obviously a poor imitation of the real thing? Indeed, why does a Real Doll so desperately try to imitate the real thing? Surely is it because a (the?) key characteristic of being human is the ability to give to that which is Other than me; humanity as intended for gift.
Back to Wittgenstein: “Why can’t my right hand give my left hand money?” he asked rhetorically in his Philosophical Investigations [though the emphasis is mine]. “The right hand can put it into my left hand.” It “can write a deed of gift and my left hand a receipt.” That’s it, though. “The further practical consequences would not be those of a gift.”
The reason is that linguistically (which was where Wittgenstein’s focus lay) the idea of ‘gift’ is predicated on a deep (ontological?) sense of otherness. The right hand can’t give the left hand money because they are insufficiently other to one another.
So with ‘Real Dolls’ and, I imagine, designer babies. We become human (or, perhaps better, fulfil our humanity) by gift; conversely, by placing ourselves in essentially solipsistic relationships of the kind we have been discussion, we deny our own humanity. That’s the last suggestion in my list of family resemblances. A robot nudges toward personhood, when it can give. And it nudges towards full personhood, the kind that most of us humans don’t even come close to, when gift becomes sacrifice, not just giving what I have but giving what I am. Think of the robot that possessed some of those family resemblance above – narrative, ethical orientation, awareness of death, desire to reproduce – and then was also willing to sacrifice itself for another robot.
One very last remark: I am struck that for all that Christian theology talks about God the Father acting before the creation of the world, and of the pre–existent Logos, and of the Trinity, the Christian scriptures are not especially interested in talking about any of this. Rather, they begin, bluntly, “In the beginning God created…” In Genesis, we don’t know anything of God before we know he is a creator, a giver of gifts that are ‘other’ to him.
Similarly, the fullest revelation of God in Christian thought comes in the kenotic Son, the one who “made himself nothing by taking the very nature of a servant”, and, more precisely, in the Son who emptied himself as utterly as is possible “by becoming obedient to death even death on a cross.”
God as gift. God as sacrifice. Here is full humanity, if we have eyes to see it. And here AI may one day travel. If it does, we should welcome the company on the journey.
Nick Spencer is Research Director at Theos | @theosnick
Image by Jiuguang Wang on Flickr available under this Creative Commons Licence
Interested by this? Share it on social media. Join our monthly e–newsletter to keep up to date with our latest research and events. And check out our Friends Programme to find out how you can help our work.