| I am thinking more and more that it’s fundamentally a Wittgensteinian kind of problem. We define a word to mean a certain collection of things (consciousness) and then try and stretch that definition to other things in the world that have the same appearance. The problem is that this abstract term likely doesn’t exist in itself as a quality, but is just a shorthand for a collection of behaviors that are observed only in biological entities. And so even if a machine exhibits all the appearance qualities of this definition of consciousness, it’s fundamentally not the same thing at all, and the only reason we think it is, is because our language is insufficient for actually describing reality. In pragmatic terms it might not actually matter, if a machine 100 years from now passes every conceivable Turing Test. But that doesn’t mean that machines have become conscious in the way humans are conscious. Expanding in an edit: It just means that the word consciousness is more descriptive, like awareness, and not the soul-derived concept that it still functions as today. Side note – I spent a couple months last year researching the history of consciousness for an essay contest, and one conclusion is how consciousness is descended in large part from our concepts of the soul. Which explains a lot of the reason it has such cultural prestige today. |
And currently its widespread usage in how people relate to and talk about Ai is actively harmful.
And dont get me started on how terrible the "hard problem" is, yeesh.