tunities. So the people that you've chosen to teach your children don't believe that AI will significantly affect their employment opportunities. What do you even say about that? Did we intentionally find the dumbest people to become teachers? How in the world could they have that opinion that the job that's most likely to disappear entirely is their job? Am I wrong? The most likely thing to disappear entirely is teachers.

Now to me it seems like the model of the future is somewhat obvious. Getting a spam call, the model that would work is obvious. You get rid of bad teachers and you replace them with good AI. Now the AI might look like people but it'll be on a screen. It might be VR someday. But in the short run, the best AI that learns from all the teachers that there are and puts together a class that is the best of the best.

But let's say you could also change who the teacher is just by pushing a button. So let's say I think it will hold my attention better if the teacher is an attractive female, so I just choose that option because I think it will keep me sitting there looking at it longer. Why not? You might pick the man with a deep voice because you're sexist and you just think that sounds more authoritative, so maybe you'd learn better in that case. Maybe if you're a member of some minority group you might say, you know, I just want somebody who looks like me to teach me. It makes me feel more comfortable. Good, you have the option. Whatever works for you.

So the fact that AI could be instantly modified for the specific person, and then let's say that you say to the AI, you know what, the way I learn best is when you put it in a story form, or the way I learn best is when you put it into a practical real world example, or I need to hear it or I need to see it or I need to write it down. The AI could work with you to figure out what your learning process is and where you need a little extra and whether it needs to repeat. It could look at your eyeballs. The AI teacher could be looking at your eyeballs as you go because if you have a camera and it could determine if it's working or not.

Imagine the AI seeing that you're starting to zone out a little bit and so it says, oh I got to change pace. I got to mix it up. I got to excite you. I got to give you a little shock. Your teacher can't do that because they're teaching to one person or they're teaching to the whole class, so they can't adjust for any one person. But AI could. They could adjust instantly to one person.

So I don't think there's any chance that in the long run teachers will do what teachers have been doing now. They might have great opportunities, but you know they might be the person running the AI or managing the process or overseeing that the classes are right. There might be something like that. But it's not going to look the same, that's for sure.

Over in New South Wales they had a program to predict crime. They were going to use AI plus state surveillance stuff to figure out who is likely to be a criminal so they could watch them a little extra. Now I have mixed feelings about this. Number one, could you build a technology with current technology that would accurately determine who is likely to commit a crime? What do you think? Do you think our current technology can identify people who are likely to commit a crime? Of course it can. Of course it can. Do you know why? Because I could do that. I could just read your damn messages and I'll tell you if you're going to commit a crime. Bet I get it right 80 percent of the time. AI could do a little better than that. Yeah, absolutely.

People totally broadcast their criminal intentions in their words and actions, very much so. It would be obvious to the AI. But we can't really live in a world in which every part of us is surveilled and every bad behavior is caught. Because the problem is that we're all bad behavior people. Not you. I mean you're the exception. And I'm the exception. But amazingly only