Quote... "Computers will overtake humans with AI at some point within the next 100 years. When that happens, we need to make sure the computers have goals alig

Quote... "Computers will overtake humans with AI at some point within the next 100 years. When that happens, we need to make sure the computers have goals aligned with ours." There has been a real bandwagon recently for people to assume that AI will automatically become dangerous as computing power increases. I'm not convinced. To be dangerous I would argue that the AI needs to be self aware and capable of creating its own motivations. Nothing in machine learning, or AI in general, has provided any insight in how to achieve this or how we might even get closer to it. A faster computer does not create either. We have much better chess computers now that can beat any human player, but they have no relevance to general intelligence. I would also argue that any emergent self aware and motivated intelligence would likely not have direct access to its own computing substrate anyway. We have consciousness that emerges from our own biological substrate but I cannot directly access and program my neurons. How come my billions of neurons are not directly programmable by me as if they were a graphics card with massively parallel power? An intelligent machine would not directly be able to use its underlying CPU directly to perform fast math, it would need to learn it as a concept and work it out as we do. The irony is that if we create an immensely powerful computer and a conscious entity emerges, it would need to learn how to do long addition they same way my 6 year old boy does.