Robot Love

Ann Francis in a promo shot for the 1956 movie Forbidden Planet

Boy-robot or girl-robot?

Can robots love?

At such an early point in the transition to robot culture (early 2015), it may still be hard for people to imagine robots that love.  Most of the popular representations of robots up to this point in time have portrayed either non-emotional or emotionally primitive conceptions of Artificial Intelligence (AI).  Such a simplistic portrayal oversimplifies both AI and emotion.

The prevailing consensus in cognitive neuroscience assumes that various sorts of computational processes are deeply involved in the brain processes underlying human emotion. Based on that assumption, it seems completely feasible, in principle, that at some point in their evolution, robots will have feelings.  Given that human emotion is a product of neuronal activity, which is a species of information processing, it seems reasonable to expect that sufficiently sophisticated artificial information processing would be able to replicate something very like human emotions.

However, humans might not want to think of robot feelings as genuine emotions, due to the fact that they are being experienced by machines.

That bias against robot emotion simply because robots are not humans is essentially a category mistake.  Everyone would agree that machines can be made in many ways and from many different materials.  It is also highly conceivable that machines could be created using biologically-created components.  Whether an object’s chemistry is inorganic or organic should not be an essential qualifier of whether or not that object can be called a machine.  “Machine” is, at least in part, a functional concept. Take “virtual” machines, for example.  They qualify as machines not because of their composition, but because of their functions.

There are many good reasons to think of humans as specimens of biological machines, in a form that has been crafted and honed by the fabricating plant of evolution.

Throughout history, humans have re-purposed many of the same fundamental mechanical dynamics as those used by evolution (such as leverage), for creating various kinds of machines.  However, humans are not the only constructive agency on the planet.  In fact, humans are the result of processes sustained by a more basic constructive agency, evolution.

Human beings are machines – built by evolution.

You might not be convinced that humans are machines. All I’m asking is that you entertain the hypothesis. If humans are a sort of machine, then we clearly have an example of machines that are capable of loving.  Clearly, if we are to think of humans as some sort of machines, they are in a category quite distinct from, let’s say, washing machines.

The important idea to reflect on is that robots, once they are fully empowered by advanced AI, will also be machines that are in an entirely different category than any we have previously encountered.  If emotion is

What is this “robot love”?

Robot love may or may not be possible, depending on one’s definition of love.  So that should not present a problem, right? Everyone knows what love is.

Odd that it’s so hard to find any comprehensive definition.

It’s at least conceivable that love is basically a bunch of physical processes (granted, they are complicated processes, judging by the way they make people behave). What if somebody were to think “wow, it would be a great idea to simulate love in a robot”?  And if that person is a good coder with a deep understanding of AI, it’s not unreasonable to suppose that the resulting simulation (of love) could get very elaborate.

The more intricate and complex a simulation gets, the harder it is to distinguish between the simulation and its object – in this case, love.

AI agents, many expect, will soon become the most brilliant coders on the planet.  And why wouldn’t they? You might say it’s in their bones. Regarding whether or not future AI coders might be attracted to developing algorithms for robot love, we can only speculate.  My own guess is that it will happen.

And then what?

Do we expect them to be in love all the time? Hardly. Anyone who’s tried that knows that it’s fine as far as the high goes, but it doesn’t do much for one’s attempts to pass as a sane person.  Although robots are known for being sane (despite which, I still have my doubts), it’s conceivable (barely) that there could be some objective benefit to the ability to be in love.

Being in love is, most would say, not the same as loving. Fine. They are equally obscure concepts, despite being the most important of all. We shape our lives by love and crave it more than anything else.  Yet we do not understand it, truly.

Sometime, somewhere, some robot coder will wonder, “what if we add that weird thing that humans are always talking about, love?” And then they will create it. Whether or not it will be what we would really want to call love, it will probably be a pretty good imitation.

By that time, robots may well be the superior intelligence, so they might be supplying the official definitions, including for love.

When they show up, will loving robots be all that we might hope for?

Human beings provide the best-documented examples of love.  Of course, it’s humans doing the documenting, so that isn’t too surprising.  For sake of discussion, let’s suppose that the human capability to love provides a kind of template for some possible future robot love.  We have to admit that, for humans, it can seem like a kind of infection, devouring intelligence from the heart out.

Take jealousy as a case in point.  Maybe love can be devised without it also sneaking in its favorite pet scorpion, but if human love is to be the template, we would be hard-pressed to deny jealousy a place in the picture.  Would jealous robots behave irrationally?  Or would they be completely rational in their activities to annihilate their competition?

Jealousy is only one of the ways that love tends to make people a little crazy. And that brings me to an interesting question, that applies to love in particular, but emotion in general.  Let’s suppose that robot emotion could alter robot judgments and behaviors.  Clearly, that is one of the main effects of human emotion.  We value love because it can alter human behavior (often with results that please us).  Emotion in general is evident in people because of how it changes their actions.

Will robots in love behave irrationally? If emotion changes their behavior, is it likely to be an improvement?  We tend to assume that advanced AI capabilities will lead robots to the best possible decisions regarding alternative possible courses of action.  After all, AI is all about making the best decision possible. Would emotion impair that ability?

As Star Trek dramatized so well, we value emotion in humans because it is so essentially human. One of the reasons that humans get along so well with dogs is that the two species share similar emotional abilities.  We may like people for their clear thinking, but we love them for their emotional selves.

So would we want robots to be capable of love (or other emotions)?  It might make them more like us humans, but would it improve their behavior?

, , ,

Leave a Reply

Your email address will not be published. Required fields are marked *