Elon Musk keeps saying that AI will destroy us and it should be regulated now. How exactly would AI destroy us?

Let's define what we mean by “destroy". It almost certainly won't cause humanity to go extinct. But, just like when people say “smartphones are destroying America” or “social media is destroying young people”, what they mean is “the effect it's having on America/young people is severely deleterious”. So, with that in mind, what deleterious effects might AI have? Well: Easy to fake porn of real, non-consenting people. Or worse, real children Loss of employment for the people replaced by AI Yes, I recognize that AI productivity should increase productivity per capita - as it's additional productivity without an additional capita - and therefore, life should be easier. However, the distribution of productivity amongst capitas is a social problem, and one we might not solve in time. Related to the last point, there's whatever social effects happen when median work week is 0, and people just have 168 hours of leisure a week. Which is mostly scary due to being unknown. Yes, the 40 hr workweek in its current form is recent, but it's never been the case that the majority of the population is secure for life with no effort needed on their part. The unknown. There's some debate about how major a technological shift would come from AI, but even a small shift has the potential for unpredicted consequences. Nobody knows exactly what AI will or won't do, and unknowns are scary. Now, will that count as “destroy”? Will there be benefits to outweigh the side effects? That remains to be seen. But perhaps that's what Elon meant - “AI could destroy us”