2 ms·
I think the basic question -- what would it take -- is actually quite simple if you unpack it. But the question conflates two totally separate things -- being
by crazygringo 4y ago
I think the basic question -- what would it take -- is actually quite simple if you unpack it.
But the question conflates two totally separate things -- being conscious and thinking.
The easy answer is the one to "thinking". And this requires it to contain an actual working mental model of the world that gets updated, and that it uses to reason and act in order to satisfy goals. This is GAI -- general artificial intelligence. And it's opposed to just the pattern recognition and habit/reflex "autocomplete" AI of something like ChatGPT. There are lots of tests you could come up with for this, the exact one doesn't really matter. And obviously there are degrees of sophistication as well, just as humans and animals vary in degrees of intelligence.
As for actual "consciousness", that's more of a question of qualia, does the AI "feel", does it have "experiences" beyond mechanical information processing. And we can't even begin to answer that for AI because we can't even answer it objectively for people or animals like dogs or dolphins or ants or things like bacteria or plants. We don't have the slightest idea what creates consciousness or how to define it beyond a subjective and often inconsistent "I know it when I see it", although there's no shortage of speculations.
As for the rest of the question -- philosophers have come up with lots of ethical frameworks, but people legitimately disagree over ethics and academic philosophers would be out of jobs if they all agreed with each other as well. When we do come up with a thinking AI, expect it to be the subject of tons of debate over ethics. And don't ever expect a consensus, although for practical reasons we'll have to eventually come to mainstream decisions in academia and law, much the same as there are for ethics in animal and human experiments currently for example.