True A.I. as possible salvation from human instinct

Abstract

In the past few years, we’ve heard a variety of philosophers and scientists proclaiming the dangers of the rise of self-aware Artificial Intelligence, for all intents and purposes. To an extent, such proclamations are rational, for if true A.I. were ever invented, it would conceivably involve some variation or another of one or a set of self-learning and self-improving basic mathematical formulae, likely to take the form of evolutionary algorithms. These would conceivably be capable of quickly evolving beyond the human ability to match them. Relatively basic versions of these, which currently serve as proofs of concept, are already in use in everyday life. From fairly recent changes in the workings of Google Search, and Google’s more recent revolutionary breakthrough called AlphaGo, to improvements to Apple’s ‘Siri’ functionality, we encounter examples of programs that contain self-modifying code which tries to mimic the learning methods of humans, specifically of human children. Our increasing insight into A.I. has led to the quick proliferation of such code, and as a result, ethical speculations about its dangers have already started to spread to the general public.