On Computability 1 – 785 views
AI is increasingly capable to simulate human behaviour and reasoning in different areas. The Turing test (“imitation game” as of 1950s) proposes the interrogation of the machine intelligence by answers to written questions and the assessment if the given answers are generated by a human or a machine. Are the machine answers indistinguishable from that of a human then the test would be considered as passed. From today’s perspective the interrogator would need to spread the questions over many fields of knowledge and perception. Questions concerning chess positions alone could probably reveal the machine by its ability to anticipate many more moves than a human…
But can consciousness, can a ghost be “planted” into a machine? Is it, in the end, a question of complexity and structure of an intelligent program? When approaching this question, a bunch of other questions arise: “what is computability?”, “what is determinism or non-determinism?”, “what is causality?” or “what is a continuum?”. In this blog category, we would like to address such questions but not solely at the level of narrative speculations and not without page-long computational examples. We start with some considerations on computability.
Download