Morality, Reality, and Technological Change

As a combination of my interests in poker, AI, and profit, I have been working for some years on an AI bot to play online poker. As a result, there have been numerous discussions over the past months about when and if operating such a bot is moral. There are two basic lines of argument, one about contractual obligations to the site, which I agree with, and one about deceiving other players and violating their expectations, which I disagree with, and hence have been arguing about.

In the context of this discussion, dagon.net said:

Most sites put limits on playing under different names, or opening accounts when you already have one. Likewise, they generally prohibit multiple different people playing the same account. This is tied into player expectations of identity continuity, and the idea that if you learn something about someone, it has a good chance of remaining true next time you see them.

I say that changing your account to evade a reputation you've made seems rather dodgy. As does running a bot in secret.

I replied:

I guess I see lack of identity continuity as a fundamental feature of online play. It is a natural characteristic of the online world. Yeah, it's different from the physical world, and people have trouble wrapping their head around things that are different. But if you try to create identity continuity online, you are fighting a helluva uphill battle. Nyms are cheap, that is the nature of the internet.

And Alex Kay replied that nyms being cheap had nothing to do with morality. I disagree, and I think there is some interesting discussion about morality to be had there, so here we go.

I am an atheist. Further, I am an evolutionist, that is, I believe that the miracle of intelligence and consciousness is the result solely of the process of evolution by selfish genes acting as though to maximize fitness. While there are certain high-level aspects of being human which are not for specific evolutionary reasons (appreciation of art, for example), all of the low-level intuitive modules in our brain are there for specific reasons. This has philosophical implications about the origins and nature of humanity's moral intuition.

We have a module to recognize faces and another to detect cheating (which, amusingly, fires if you phrase a situation in terms of people but not in terms of boring words and numbers - clearly demonstrating that it was meant for use in social situations). In fact, we have an astonishing variety of specific modules which are most noticed in the breach, when brain damage takes out a specific module, leaving the person, say, unable to recognize faces, yet perfectly competent in every other way.

Setting aside the lofty cerebral moral philosophy about what ought to be right and what about to be wrong (whatever that means), our moral intuitions (instinctive judgements about right and wrong) are clearly a low-level module. Therefore (if you are an atheist and an evolutionist) they must have been evolved to serve reproductive fitness. Therefore those moral intuitions do not come from God, nor from Rand, but from many generations of selfish optimization[1]. So if we are looking for some kind of justifiable and absolute morality, we had best look elsewhere, and should cast great suspicion on our moral intuitions.

Why? Well, if there is one general thing to be said about evolved intuitions, it is that they are brittle. They are trained on past data, and not updated with new data, so when the world changes, they don't change with it. We are in a time of rapid technological change, which puts great stress on all our intuitions[2].

Which is not to say that our moral intuitions are not valuable - after all, they embody millenia of experimentation and wisdom about how people can get along in groups. Just like I consider it an amazing and wonderful thing that the blind selfishness of evolution has created intelligent life, I consider it an amazing and wonderful thing that the blind selfishness of evolution has endowed us with the capacity for trust, honor, honesty, and nobility. But we must not let our appreciation make us forget the base origins of our moral intuition, which were trained on old data and built to serve our genes[3].

Now we get to the connection between moral intuition and reality. Moral intuition is a mechanism evolved to help people help themselves by getting along with each other. And the ways in which people can cooperate, share resources, borrow and lend, are fundamentally tied to reality. Therefore moral intuitions are based on reality, and so facts like "nyms are cheap" can affect morality. Unfortunately, they tend to affect it slowly, through evolution, although they can also affect it more quickly, through culture[4], which takes only one generation. This is rather abstract, so let me give you an example:

David Friedman suggests (in Law's Order, I think) that private property in land may have come into existence with the domestication of the dog. In other words, the morality followed from the technology. Now, I have a very strong moral intuition that my land is my goddamn land, and you better not try to take it. Yet before the technology existed to privatize land, I don't believe humans had this intuition. It's not like people were complaining about how other people kept using their land for hundreds of thousands of years until the dog. The very concept of private land arose along with the technology to protect it.

Sexual mores is another good example. Human morality has a lot of (sometimes conflicting) things to say about sex, and a little examination will show you that it's beliefs are firmly rooted in genetic practicality. Male anger over female infidelity, for example, is because it is a genetic disaster for a man to waste resources raising another man's child. Yet women sometimes cheat because the qualities of a good provider and an alpha male are different, and if she can have the genes of the latter and the resources of the former, it is a huge genetic win.

As David Friedman points out, birth control, abortions and paternity tests destroy the basis for this part of morality. It is no longer necessary to jealously guard your wife from screwing around, because the reason for doing so is obsolete. Yet, because this morality is hardwired into us, most people still try for monogamy. Many people believe that having sex with someone other than your spouse is wrong, even if you are honest about it, and not because of contractual obligations, but because it's just wrong. Yet if the current state of technology continued for many generations, that morality would change, because the selection pressure for it is gone. (Just like primates with different reproductive incentives have different sexual mores.) It is a moral lame duck, orphaned by the changes in technology.

Now we move to the online world. Yes, we are used to continuity of identity. But that is because in the physical world, that is the only way things can work. In the online world, there are many more strangers, and so you should trust people less. It is a fundamental part of modern reality that people can easily make online nyms, and if morality had time to evolve, it would surely take that into account. We would intuit that people online could be anyone, that we should be reluctant to trust them, that when you are arguing with or playing poker with someone online, for all you know you are in a battle of wits against the latest computer program.

To sum up: Morality is not independent of reality, it is fundamentally based on it. And in an age of rapid technological change, we must be cautious of intuitions trained on out-of-date data.

[1] The book that best conveys this is The Origins of Virtue, by Matt Ridley.

[2] This is discussed extensively, although in a legal, not moral context, in David Friedman's Future Imperfect.

[3] To learn about the general problem of being a robot programmed to serve interests other than your own, which I believe is a fundamental part of the nature of being human, see The Robot's Rebellion.

[4] Note that from this perspective, culture is a very useful tool, because it is how we can change morality without waiting for genes to catch up. The rebellion of each generation against the next is actually damn useful, and so the moral beliefs of children are an important clue about future cultural attitudes.

Share this