Do AI have sex? And why they should
Updated: Feb 10, 2019
What if our mind is just an (or THE) ornamental feature for sexual selection, as the beak of a toucan or the tail of a peacock? For some evolutionary biologists, and certainly for Geoffrey Miller, this is the case. For any evolutionary biologist, there are two different mechanisms for gradual changing and improving of a species. One is the natural selection – which pertains to the development of survival features in a given environment. The other is sexual selection – that which determines the development of certain characteristics that are deemed attractive for the other sex, and make one individual more successful to reproduce.
While the sexual selection is widely discussed in the animal world, the humans did not receive the same careful examination till recently. In his book “The Mating Mind. How Sexual Choice Shaped the Evolution of Human Nature”, Miller argues that the primary function of the brain – this expensive to develop and maintain organ – was not adaptation for the natural selection, but for the sexual one. That is, the brain’s main function was, and continues to be, courtship, not creating art or innovating technology.
The argument is simple: everything that the brain created with incidentally high impact on our adaptation to survival in harsh environments (including space) is a secondary by-product of its development in wits, language, and seduction strategies aimed to secure the sexual selection and reproduction of individual genes. How? Well, simply, if we think about it, the great achievements in human culture came in the last 10 000 years, while we are roaming around with the same brain, talking to each other, and eventually talking each other into having reproductive sex back 100 000 years. Nature, the author says, innovates for immediate functionality, it does not project in the future. It did not plan to produce an organ able to develop language thinking that 90 000 years after this accummulation and transmission of experience through language this organ will enable the development of a sophisticated culture that will increase our survival rates by modifying the very environment in which we live in. In his words “Evolution does not have a Protestant ethic of work. It does not get tax credits for research and development. It cannot understand how a costly investment in big brains today may be justified by cultural riches tomorrow” (p. 20). Thus “evolution found a way to act directly on the mental sophistication of this primate species, not through some unique combination of survival challenges, but through the species setting itself a strange new game of reproduction. They started selecting one another for their brains.” (p.21).
For the sake of the argument let us accept this hypothesis as being accurate, and let us project this upon AI. How is AI suppose to learn and evolve? The models we use are still tributary to our projection of what intelligence means. And since our current approach on intelligence is one where sex and bodily manifestations are relegated to an inferior, animalistic part of humanity, almost seen as not part of humanity, let alone intelligence, I would speculate that the machines will not court each other. The question is: how will they chose with whom to mate? That is, with what other machine to network? Or to what other network to link? All these in fact greatly depend on how are we to conceive of AI. Is it a vast neuronal-like network? Than, mating is not really a question, although again, choosing with which part of the network to create synapses may be a question of taste and affinities.
If individual machines are endowed with AI, then they must independently reproduce at some point. The question is, how will hey do it? And how will they choose with whom they will do it – that is if they will be programmed to reproduce otherwise but by self-multiplication or“cloning”. At this point, AI’s sexuality, if there will be one at all, is still a big question mark. However, without it, the evolution mechanism of AI will lack an important leg. It will only be a survival of the fittest type of evolution, without the biological exuberance of the sexual evolution. Therefore, I think, machines and algorithms should have sex. And those designing them should think seriously about it.