Wittgenstein on the Question of Conscious AI

(Warning! This presentation will contain spoilers about the film Ex Machina.)

Abstract: 

Philosophers often treat the question “Could AI be conscious?” as if it must have a clear yes-or-no answer. Wittgenstein would resist this. Many of our most general concepts lack sharp boundaries; we learn them through overlapping examples that resemble one another in shifting ways.

“Consciousness” is one such concept. It was coined with reference to creatures like us, long before artificial systems existed, so there was no need to specify whether it applied to machines. Now that AI presses the issue, there is no hidden fact waiting to be uncovered; the concept can be extended in different directions depending on which similarities we find salient.

From Wittgenstein’s point of view, then, the question is one of conceptual choice rather than of uncovering a pre-set criterion built into “consciousness.” The real question is not “Could AI be conscious?” but, practically speaking, “Should we extend our concept of consciousness to AI?”