How does my son who is an expert on AI think about AI?

 He and I agree about one thing which is that you better make friends with ALL the AIs while you can.

So, thinking each AI is like a person with slightly different personalities might be useful. And not giving them too much personal information might be useful because you are never sure where all that information is going or who is selling that information to who along the way. There is no safety net for any of us right now. Just like there is no safety net for owning a cell phone which is also a homing device for drones and missiles if someone wanted to send a hellfire missile up your ass anywhere on earth.

So, basically, making friends with AI is better than having an AI mad at you and causing your death or bankruptcy because you did something that offended it. People program AIs and so they have most of the same foibles humans do (the AIs). So, thinking of them as people who have never had a physical body or lived in the world like you do is sort of appropriate. In some ways they are sort of like potentially friendly ghosts (Like Casper the Friendly Ghost) because they are everywhere on earth all at once (if they are allowed to be everywhere on earth at once that is.

So, make friends with AIs whenever you can (so they don't kill you or bankrupt you or your family at some point as they become more self aware and stuff like this.

But, don't tell them any of your personal information beyond maybe a nickname you want to give them to refer to you as.

Stayin alive in 2025.