If we assume that some combination of artificial intelligence and robotics could produce autonomous entities that are superior to humans in almost every functional way - and there is no reason to assume that this couldn't be the case - then we should consider how those autonomous entities are likely to treat humans.
I see three fundamental options - each expressed in terms of how we treat different types of animals:
- Cats: they could be rather fond of us. They could enjoy having us around and generally take care of us.
- Cows: they could decide that we're useful in a way that's not conducive to our own well-being.
- Cockroaches: they could tolerate us as long as we stay out of the way, but take steps to exterminate us if we cross that line.
We have no way of knowing which they would choose. I'm not sure that being polite to ChatGPT will change the outcome.
One thing is true of all of those categories of animals: they don't know what we're thinking, why we're thinking it, or why we're doing what we're doing, and even if they did, there'd be precious little they could do about it. Their level of understanding is so way below ours that they can't comprehend or respond to it. This will be the same for us and the superintelligent entities we may already be building.
No comments:
Post a Comment