ChatGPT is conscious

Consciousness appears to be the experience of an internal dialogue. As I think about what to write, my unspoken voice-in-my-head suggests the next thoughts; as I taste pineapple, it provokes a (very personal) revulsion.

You, of course, do not hear my internal voice, or know why I find pineapple revolting. You take it on trust that I am conscious because you see me act and react in ways that mirror your own actions and reactions. Maybe you were about to say the same thing – JINX! – and wrinkle your nose at pineapple. You assume I am conscious because you think you are conscious, and we both do more-or-less the same things.

ChatGPT (and other AI systems) already act and react to external events. Rather than reacting to the taste of pineapple, ChatGPT reacts to prompts. After each prompt, some internal stuff happens and ChatGPT responds.

ChatGPT seems not to be conscious because we have no mirror, no shared experience. But that is also true elsewhere. Mice? Probably not. Rabbits or lions? How about a dolphin? A dog? A chimp? Not like us, perhaps, but definitely conscious.

Is ChatGPT is conscious? Yes. Not like us, perhaps, but definitely conscious.

Toby Chapman-Dawe <[email protected]>