Artificial life

More than 60 years ago, the young Irish crystallographer John Bernal anticipated the possibility of machines with a life-like ability to reproduce. "To make life itself will be only a preliminary stage," he wrote. "The mere making of life would only be important if we intended to allow it to evolve of itself anew."

Almost two decades later, John von Neumann performed the first work that suggested the possibility of artificial life, in his studies of the logic of self-reproduction and its relationship to complexity. Efforts to put (metaphorical) flesh and bones on his ideas followed. In 1956, a proposal was outlined for "artificial living planets", self-reproducing floating factories that could harvest important mineral and crop resources. Recognising the dangers posed if such a machine ran amok on Earth, Freeman Dyson formulated a more benign suggestion in which self-reproducing machines seeded life in the solar system, away from the Earth.

Today, however, von Neumann's insight into the logical nature of life has more significance than ever because it is now possible to perform Darwinian evolution by "natural selection" inside a computer. In the same way that biological life ultimately emerges from the complex interactions between a great number of microscopic units called molecules, so it seems that artificial life (ALife) may emerge from the complex logical interactions between bits within a computer.

In a certain sense, ALife research has been under way for decades, albeit by a different name. Computers have been used to simulate a wide range of biological processes, by solving a specified set of equations that are believed to model the phenomenon under study, such as the pattern of limb growth or aggregation of a slime mold. Just as a computer model of a nuclear explosion is not itself a nuclear explosion, so these kind of biological simulations are in no way alive. Those who pursue such research can be called supporters of "weak" ALife (by analogy with supporters of "weak" artificial intelligence (AI)), since they study computer models of biological processes in which the simulations could not be termed living.

What distinguishes a certain strand of contemporary research in the field from the mainstream of study is an often unspoken belief in "strong" ALife, according to which a suitably programmed computer may itself be deemed to be alive, or at least possesses properties of a living thing (cf. "strong" AI). Since it is difficult to find two people who can agree on a definition of life, one might think that attempting to define artificial life would only increase the confusion. Petty arguments about definitions have not discouraged the hubristic claims of the strong ALife community.

One of the most provocative was made by Doyne Farmer and Alletta Belin: "Within 50 to 100 years, a new class of organism is likely to emergeIThe advent of artificial life will be the most significant historical event since the emergence of human beings. The impact on humanity and the biosphere could be enormousI" Although there are overtones here of the claims made by supporters of strong AI for the emergence of intelligent machines, which have so far ignominiously failed to materialise, the philosopher Elliot Sober has observed that terrestrial life is in many ways far better understood than the human brain, so there is a firmer grounding for ALife than for AI.