Saturday, November 27, 2004

Artificial Life

From a paper on Artificial Life I read recently:
Ideally, the science of biology should embrace all forms of life. However, in practice, it has been restricted to the study of a single instance of life, life on earth. Because biology is based on a sample size of one, we cannot know what features of life are peculiar to earth, and what features are general, characteristic of all life. A truly comparative natural biology would require interplanetary travel, which is light-years away. The ideal experimental evolutionary biology would involve creation of multiple planetary systems, some essentially identical, others varying by a parameter of interest, and observing them for billions of years. A practical alternative to an inter-planetary or mythical biology is to create synthetic life in a computer. 'Evolution in a bottle' provides a valuable tool for the experimental study of evolution and ecology.

-- Thomas Ray, 'An Approach to the Synthesis of Life', in M. Boden (ed.), The Philosophy of Artificial Life, p.111.

In that same anthology, Christopher Langton characterises traditional biology as an analytic endeavour, starting with the whole organism and breaking it down into its component parts. Artificial Life takes the opposite (synthetic) approach, trying to "put living things together".

One of A-Life's core concepts is that of emergence: complicated global behaviour can arise from simple local behaviour. For example, artificial "Boids" have been invented which imitate the flocking behaviour of birds and fish. The behaviour of individual Boids is based on three simple rules (e.g. "move toward the average position of local flockmates"). But a group of Boids will exhibit quite complicated 'flocking' behaviour, such as breaking into sub-flocks to flow around an obstacle, then seamlessly merging together again afterwards.

More interesting A-Life research involves virtual populations that evolve. Ray's Tierra system is a remarkable example of this. He started off with a basic self-replicating program, and set it in a virtual environment which allowed random 'mutations' to arise by way of occasional copying errors and such. All sorts of fascinating new programs evolved, simply through these random mutations and the 'natural selection' of having to compete for CPU time in order to reproduce. Parasites arose which were much smaller because they had no 'reproductive' code of their own, but instead found a full program and hijacked its reproductive routines! In light of this new selection pressure, new programs evolved which were immune to the parasites, or could even utilise the parasites to their own ends (and eventually drive them to extinction).

Fascinating stuff, I reckon. A simple introduction to A-Life, by the way, can be found here. If you want to see the power of selection for yourself, try out the fun (if slightly silly) Biomorphs game, where you select which of several randomly-mutated 'organisms' makes it into each successive generation. You can quickly 'evolve' something that looks like a bird, or a person, or pretty much whatever you want.

An interesting philosophical question which arises from all this is: Could a computer program ever be truly alive? Ray says, "I would consider a system to be living if it is self-replicating, and capable of open-ended evolution." His Tierra programs qualify, by those criteria.

It seems odd to call a computer program 'alive', but then again, there seems little reason for us to insist that life must be carbon-based. As Langton points out (Boden, p.53):
Life is a property of form, not matter, a result of the organization of matter rather than something that inheres in the matter itself. Neither nucleotides nor amino acids nor any other carbon-chain molecule is alive - yet put them together in the right way, and the dynamic behaviour that emerges out of their interactions is what we call life. It is effects, not things, upon which life is based - life is a kind of behaviour, not a kind of stuff - and as such, it is constituted of simpler behaviours, not simpler stuff.

So, if a computer program can exhibit all the behaviour we associate with living organisms, perhaps we really ought to say that it too is alive?

Update: Carl Zimmer has a great article on the topic.

2 comments:

  1. a curious problem arises - if a man can produce a computer program that either is more intelligent than any human OR feels more pain than any human - should we immediatly consider ourselves lower down the scale of importance than that creation and since such an invention is basically inevitable does that mean that we will eventually from a moral perspective have to sacrifice ourselves for the benefit of a higher organism that we create? possibly even a lower organizm that jsut happens to fit the moral criteria (such as pain) better than we do.
    For example imagine if such an organism requires a large amount of entergy to survive and wishes to reproduce etc. and as a higer organism than us demands those as moral rights. 

    Posted by geniusNZ

    ReplyDelete
  2. Interesting questions. Though the creation of a conscious (pain-feeling) A.I. is far from "inevitable" - indeed, whether it is even possible is a highly contentious issue. (I think it is, but a lot of very smart people disagree.)

    But the general moral question here is an interesting one: Would humans have a greater moral obligation to any "superior" species we were to come across? Or should we always look after ourselves first? I'm not sure. 

    Posted by Richard

    ReplyDelete

Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)