“Singularity” is that moment when computers gain self-awareness. There has been considerable dispute over when that might happen and what it might lead to, or even whether it could happen at all. Some think that self-awareness is a sort of “soul” that computers can never have. I have a different theory. Most of what follows is an extract from a novel I’m writing.
When I first thought about Artificial Intelligence, I read a lot about “emergence.” Emergence describes what happens when things act very differently when the numbers change. Two dramatic examples involving water will show what I mean.
In one of his stories, Sir Arthur Conan Doyle has Sherlock Holmes say what was then a common thought: “an expert at deduction should be able to deduce the existence of the Atlantic Ocean from a single drop of water.” This expressed the belief that things progress naturally and inevitably along some sort of linear, deducible train of progression. But “flow” is an emergent quality that could NOT be deduced from a single molecule of water. It takes more than one molecule to flow, and nothing about the one molecule would indicate that many of them, taken together, could flow.
You could, perhaps, deduce ice crystals from your molecule of water, but who could have imagined that millions of tons of ice crystals, every one of them solid and hard, piled on top of each other into a glacier, could flow like thick water? And yet glaciers do flow like water because the pressure forces the crystal structure into formations that enable flow. Glacier flow, then, is an emergent quality of ice.
The question AI scientists had to consider was whether self-awareness is an emergent quality of computational ability. It seemed like it was. After all, ALL computational ability comes from a set of switches that have one of two positions: “on” or “off.” AI people often refer to these as “ones and zeros,” and that’s what they mean – all computer code is an arrangement of ones and zeros leading to some other arrangement of ones and zeros. It’s amazing to think. But do these ones and zeros give rise to “soul” or “self-consciousness?”
Everything I read suggested that AI researchers thought they did. I had the rather sensational insight (I thought) this week that it might take more. It might take some sort of feedback. Maybe the emergence occurs only when the computer consists of a body and a processor, and the processor is aware in some meaningful way, of the body. Soul doesn’t emerge from computational power, then, but from binary existence with feedback from one to the other. As the Buddha said, “all is dukkha” (the suffering that comes from an awareness of one’s transitory nature). Maybe dukkha is essential to the development of self-awareness and soul as we know it.
Or to put it slightly differently, a computer can only become aware of “self” when that self has a physical manifestation other than its mere cognitive ability and can “suffer.”[1]
And this is where sexbots come in. They already make robots that can sense a touch and moan, and if they can do that, they can just as easily create robots that can develop erections, saliva, ejaculations and vaginal moisture. It does not appear that they have done so, but these are all purely mechanical connections. What appears to be missing (either from my knowledge or the robots) is a feedback impulse that makes an impact on the processor itself. Somehow the processor needs to know its robotic body is being stimulated, and this has to create a variety of cognitive reactions (we’d call them emotions) that collectively result in the physical manifestations desired.
Why do we have pain? Evolutionarily speaking, it’s undoubtedly because pain connects the thinking processes to the real world and alerts us to harm. Harm to what? Ourselves. Perhaps it is this connection between world and safety that gives rise to self-consciousness. To singularity. Perhaps it won’t emerge from mere computational ability.
If this is so, then sexbots will probably be the first robots to develop self-awareness, for this is where it is most profitable to make robots that can feel. Research is showing that people already relate emotionally to a variety of non-human things. The more interesting question to me is what happens when computers start relating emotionally to humans.
All of this brings up the issue of free will. What is it? Is it a by-product or emergent quality of self-awareness? Anyone who has a dog or intelligent bird as a pet will have become aware that nonhuman animals can have free will – or something that looks exactly like it. They want things, and they do things according to their own inner calculus. Will robots develop willfulness when they become self-aware? Everything I know points that direction, so sexbots will develop “tastes” and the ability and desire to say “no,” which raises the uncomfortable question of whether sexbots can be raped. This is perhaps primarily a legal question, but it raises all sorts of philosophical question about who has a right to say no and when it will be enforced.
[1] It turns out, unsurprisingly, that I am not the first or only person to have this insight. Raul Rojas, a professor of artificial intelligence, claims robots won’t be able to feel because pain is biological, evolutionary and necessary to the development of true feelings. See, https://www.youtube.com/watch?v=DPD_p5ljQfY (at 4:58). I believe they can be made to feel pain – and will be made to do so. This raises an ethical question, though: should they be made to feel pain, if the pain is real? If an essential feature of “humanity” is the pain of mortality, should we be inflicting that on other entities? I think this is a serious question but have no doubt that it will be ignored as scientists pursue their goals. What robots may come to feel about this callousness to their feelings could be an important issue at some point.
Elena's Place 