I always thought that the paper clip problem fundamentally missed the point. In order for the scenario to be realistic the AI would have to be super intelligent, otherwise we would just switch it off. If it’s super intelligent, surely it understands why converting the entire planet into paper clips would be a bad thing to do.
So it’s either stupid enough to actually try it, which means it’s stupid enough for us to be able to defeat, or it’s intelligent enough that we can’t defeat it, which means it’s intelligent enough not to do it. Either way the world remains unpaper clipped.
The problem is that an AI built to maximize paperclips might conclude that converting the planet to paperclips is an acceptable cost of maximizing paperclip production. It might understand why humans think it’s bad to convert the planet, but disagree. It would need to be explicitly programmed to prioritize human life over paperclips.
otherwise we would just switch it off
If it were super-intelligent, it could probably trick us into leaving it turned on.
I always thought that the paper clip problem fundamentally missed the point. In order for the scenario to be realistic the AI would have to be super intelligent, otherwise we would just switch it off. If it’s super intelligent, surely it understands why converting the entire planet into paper clips would be a bad thing to do.
So it’s either stupid enough to actually try it, which means it’s stupid enough for us to be able to defeat, or it’s intelligent enough that we can’t defeat it, which means it’s intelligent enough not to do it. Either way the world remains unpaper clipped.
The problem is that an AI built to maximize paperclips might conclude that converting the planet to paperclips is an acceptable cost of maximizing paperclip production. It might understand why humans think it’s bad to convert the planet, but disagree. It would need to be explicitly programmed to prioritize human life over paperclips.
If it were super-intelligent, it could probably trick us into leaving it turned on.