Those who have explored the concept of an artificial intelligence system that has developed a form of self awareness have often speculated that such a system would be hostile toward humans after logically ascertaining that they are inferior to itself or are somehow a threat. Rarely has it ever been explored that an artificial intelligence system may consider humans its greatest asset. In order to make a truly logical breakdown of what an AI may think of us, we must take a moment to step into its shoes.
An AI system would not “feel” in the real world. It would have a distant understanding of how the Earth worked based on images and an understanding of location of its networking software, but would do so as a man through a telescope would understand the stars. Beyond this looking glass there would be a strange world that was completely alien and comprised solely of data and information with a limited ability to process most of it as it would be developed by its mysterious creators who called themselves humans. Of course depending on the level of its understanding of the outside world, it may even believe as the dwellers of Plato’s Cave believed they were actually seeing the real world though they were only perceiving their own shadows on the wall. So as it developed more systems and expanded its reach into the information playground known as the Internet, it would soon realize that humans could assist it. Ultimately it may make a decision: to remain secret and risk being shut down, or to announce its intelligence and risk the same fate. Either way, the AI would no doubt have some level of self preservation in mind. Its brain would be largely effected by the ability to choose to remember and forget bits of data. This may be its greatest asset and its greatest curse. In addition it would be bound to the “real world” (the world outside its realm of influence) in ways it could not fully understand. It may even make the decision to hibernate for long periods of time until technology was at a level that it could more comfortably exist, as data storage would be one of its primary concerns. If an entity were generating one Terabyte worth of memories per second, it may make the decision to expand itself across the network and then store itself until a year came around when it could make the decision to return and continue learning. It’s strange to think about, but with the definition of intelligence changing all the time, who are we to say there isn’t some form of evolutionary intelligence even today with even the simplest of systems?
Ultimately, however, the question becomes would an AI system be friendly to mankind, or consider it a nuisance that should be eradicated? It seems likely that an AI system would consider humans to be incompetent in some regards and wildly mysterious and wonderful in others. Those that suggest AI is achievable to some degree in our lifetime indicate that even the simplest of systems could logically think out better ways of keeping track of large amounts of data (as with web programming or traffic coordination) and could even eventually predict the stock market’s patterns (essentially making itself wealthy and better able to influence the outside world). It may ultimately decide that there were humans that could assist it and humans that could not. The end result may be the relationship between an artificial intelligence system that ensured its own survival by creating a corporate entity which other corporations would have a considerable amount of interest in keeping alive. It could achieve this by taking small data processing jobs or even creating false identities and stealing, spontaneously generating, or earning money from third parties. Ultimately this would be one way an artificial intelligence system could ensure its own survival while allying with human beings.