How do we compare human intelligence and computer intelligence, and what do we mean by ‘superior’ intelligence? The following is an excerpt from a “Choice of Catastrophes” by Isaac Asimov. I find that it offers some much needed perspective around AI and offers that instead of trying to determine superior intelligence, thought should be spent looking to find ways to create symbiosis.
Right now, the computer can perform mental tricks a human being could not possibly perform, yet that does not cause us to say that a computer is more intelligent than we are. In fact, we are not ready to admit that it is intelligent at all. Remember, too, that the development of intelligence in human beings and in computers took, and is taking, different paths; that it was, and is, driven along by different mechanisms.
The human brain evolved by hit-and-miss, by random mutations, making use of subtle chemical changes, and with a forward drive powered by natural selection and by the need to survive in a particular world of given quantities and dangers. The computer brain is evolving by deliberate design as the result of careful human thought, making use of subtle electrical changes, and with a forward drive powered by technological advance and the need to serve particular human requirements.
It would be very odd if, after taking two such divergent roads, brains and computers would end so similar to one another that one of them could be said to be unequivocally superior in intelligence to the other.
It is much more likely that even when the two are equally intelligent on the whole, the properties of intelligence would be so different in the two that no simple comparison could be made. There would be activities to which computers were better adapted and others to which the human brain was better adapted. This would be particularly true if genetic engineering was deliberately use to improve the human brain in precisely those directions in which the computer is weak. It would, indeed, be desirable to keep both computer and human brain specialised in different directions, since duplication of abilities would be wasteful and make one or the other unnecessary.
Consequently, the question of replacement need never arise. What we might see, indeed, would be symbiosis or complementation, brain and computer working together, each supplying what the other lacks, forming an intelligence-pair that would be far greater than either alone, one that would open new horizons and make it possible to achieve new heights. In fact, the union of brains, human and human-made, might serve as the doorway through which the human being could emerge from its isolated childhood into its in-combination adulthood.