LogFAQs > #922478596

Is AI dangerous?

That does seem to be a general concern. The problem is that AI can be used to create robots that are dangerous. It can't be used to build self-replicating machines that might not even be dangerousthey're safe.

I think of AI as a weapon or tool, and the problem is that that weapon or tool is not only being used to killit's being used by others to kill themselves. In this way, the danger is not limited to AI itself, I think, but the idea that machine intelligence can be used to do bad things by using it too often. This is not a new way of thinking about AI, which is why it seemed so frightening in the early dayswe talked about a time when the military was looking for machines that could think and act in ways that might prevent wars, for example. And we saw these technologies being used to build weapons, which were used in an even way, to create new things.

Advertisement

It's really scary to see something like "Halloween Robot" come to life. What's your reaction?

I think the real scary factor might come from the fact that, in this current context, our fear of AI is not the use of the technology, but instead of the use of the technology, even though we'd like to think that the use of AI