Man Versus Machine: Grappling with Analogy

There have been some remarkable advances made with machine learning in recent years. The abilities of artificial intelligence (AI) technology are often evaluated using the Turing test, which leads to terminology normally applied to human beings being applied to AI. After a brief introduction to machine learning, this paper notes that this approach to evaluating AI is analogical and argues that the analogy is not a stable one. The most obvious and used analogy is with the human mind, which itself presents an unresolved problem in philosophy. When the details of some of the common brain-mind theories are examined, they tend to undermine the analogical argument in a way that parallels some of John Searle's concerns.