Keeping with our recent theme of techno-paranoia, a new book by James Barratt (succinctly reviewed by Greg Scoblete of RealClearTechnology), Our Final Invention: Artificial Intelligence and the End of the Human Era, posits the reasonably imminent doom of humankind via the very machines our top scientists and engineers are clamoring to build. The concept that technology — driven by artificial intelligence (AI) — will one day exceed and ultimately turn against their human creators has troubled authors of science-fiction for far longer than scientists (see books and stories by the likes of Isaac Asimov, Arthur C. Clarke, Harlan Ellison, and many others, or films by James Cameron, Ridley Scott, the Wachowskis, and so on). This is not to diminish the apocalyptic claim — it would not be the first time life imitated art.
But what exactly do we mean when we speak of computers being “smarter” than their users? My chief skepticism surrounding any grand prognostication regarding AI matching or surpassing human intelligence is the implicit assumption that we (or at least the AI programmers) have a firm grasp of what human intelligence is. Intelligence and the processes of learning we tend to associate with it are multi-faceted constructs. While computers already have the upper hand in terms of speed and storage space by orders of magnitude, these aspects alone do not encapsulate the human experience. As AI systems grow more sophisticated and are able to learn at accelerating rates, should we regard this as a mastery of what it means to be “intelligent”? What of emotional processing, intuition, creativity, morality, and other modes of thinking and being less concretely understood but no less fundamental to the ingenuity and productivity of human society? Can there ever be, to quote Gilbert Ryle, a “ghost in the machine”, a computer with a soul? Or would it be sufficient for an AI to seem human to an outside observer, a la Searle’s Chinese Room thought experiment?
Of course, the more pertinent question from Barratt’s viewpoint may ultimately be: will that which makes us uniquely human be enough to save us from the deadly rise of AI?