Richard Smith: What will robots do when they take us over?

Richard SmithLord Rees, the Astronomer Royal and former president of the Royal Society, believes that robots might replace human beings within 50-60 years. Looking at writings from him on the web, I’m not sure that’s exactly what he believes, but an audience at the London School of Economics was told he did by his friend Lord (Melvyn) Bragg, the broadcaster and author. The idea seems to me plausible and got me wondering what the robots would do once they took over.
The idea is plausible because the capabilities of robots are increasing very rapidly, already they can beat any human at chess and do many things of which humans are incapable. As scientists begin to use biological materials in creating robots, copy the human brain, and exploit nanotechnology their capabilities, including to “think” and “plan,” could far outstrip ours. Reproducing themselves should be easy. Then they can tolerate environments that are impossible for humans, don’t need to be fed or excrete, and aren’t beset with all those humans problems of emotion, needing to be loved, aggression, lying, corruption, decay, dementia, disease, and general “lack of fitness for purpose.”

Rees hopes that we can keep them as “idiots savant,” serving us but having no aspirations of their own. But he worries as well that as they become increasingly able they will take on ambitions of their own.

Why, my wife asks, would they want to take us over? One answer might be “simply because they can.” You wouldn’t want to be a servant to your dog (any more than you are already), and why would they want to be servants to creatures so much more stupid than themselves and, worse, so prone to damaging themselves and each other? Then they might look at what humans are doing to the planet and think “this is getting so bad that we too could be destroyed, through a nuclear holocaust perhaps.”

This last hypothesis supposes that they want to survive, but would they? The question that kept me awake some of last night after hearing of Rees’s worry is “What exactly would robots want to do?”

I started by thinking about what we humans want to do? Viewed in evolutionary terms, our task as individuals is to reproduce and get out of the way. As a species we have to keep going, allowing evolution to continue so that we can evolve into something much better than we are now. Another of Rees’s saying that has always fascinated me is that the “creatures” who will see the end of our planet in five billion years’ time will be as different from us as we are from amoebae. And, perhaps putting his two ideas together, evolution will jump from the biological to the physical.

But then I try to imagine the individual robot and relate him/her/it to me. I don’t know why I’m here, and, although I have reproduced (more than was strictly necessary), I didn’t think of that as my main task, and I’m now “through with reproduction.” I could have killed myself (or could do now that I’ve completed my evolutionary task), but I haven’t ever considered that. I feel I have to keep going, and I’ve constructed some sort of meaning to my life through relationships, work, learning, nature, and the arts, particularly reading and listening to music. I’ve stayed away from religion.
But would robots feel all this? Would they feel they “had to keep going”? Would they feel the need to reproduce, manifesting itself as sexual desire, something we call love, and an overwhelming urge to protect and cherish their “children”? Will they care about relationships, nature, music, and learning? Or as they are splendidly rational, will they decide like Sophocles that “To never have been born may be the greatest boon of all.”

If they are to be the next evolutionary step, then they will need at least collectively to feel impelled to survive. I find myself groping towards the conclusion that all those “human things,” like relationships and love, will either need to be built into the robots, making them little better than us, or some alternative will need to be found. Or maybe those robots will be the end of evolution, replacing us with all our flaws, and then deciding that continuing is not worthwhile and closing down.

Perhaps, to come back to Rees, there are many other planets out there among the nearly infinite number, where this scenario has already played out, and anybody or anything arriving on those planets will find just extinct robots.

Competing interest: RS is a human being.

Richard Smith was the editor of the BMJ until 2004 and is director of the United Health Group’s chronic disease initiative.