The Costs and Benefits of AI

I'm probably in the middle just because it is too hard for me to attribute any probabilities to the scariest scenarios. But even if we don't talk about destroying humanity, I still think that there's a set of harms that are very likely. If we think about the AI facilitating discrimination, undermining privacy, enabling large-scale surveillance, I think there are very real risks that are certain to happen and that really call for a regulatory action. And last question, how alarmed are you by all this? I know it's not your main field, but obviously anybody who's thinking about AI has heard these warnings,. including by the creators of AI, that it could destroy jobs or, in