I think it’s a super interesting question. But experts agree that the current form of AI chatbots can’t have sentience or a conciousness. It can’t learn from interactions with the world. And it doesn’t have a state of mind. And that’s the end of the debate. Mind that self-awareness and consciousness aren’t the same thing. And as far as I know the terms like “sentience” aren’t really well defined.
I agree with Geoff Hinton’s view that AI has shown exceptional creativity/performance in several narrow domains. But I very much dislike in the video/talk that he outlines a useful concept of subjectivity, and they immediately drop it and talk about sth else which isn’t even half as interesting.
What I completely disagree on is the stance on open-source. They’re missing that AI is a really powerful tool and will be a ubiquitous part of our future world. Which is where their analogy to the atomic bomb fails. And since training AI costs multi millions, it’s inevitably going to lead to a corporate dystopia where big corporations will have all the power and the people/humans are playthings. We’ve seen enough sci-fi movies about that. And is gatekeeping tech with money really a good thing? I mean there are bad companies out there. And it doesn’t necessarily need malice. One mistake by some employee is enough and we end up with Skynet from the Terminator anyways. I share however the perspective that open-sourcing AI is dangerous and will have consequences. But there’s no way around it if we want progress to lead into a nice future.
And I think we don’t have to talk about Blake Lemoine anymore. That episode is two years old. And the claims and arguments have been refuted and disproven over and over again. Don’t get me wrong, it’s certainly an interesting question. But it’s more a lesson in human psychology. Concerning AI he’s been wrong and we know that for nearly two years already.
Add comment