So I know I should, but I never read the terms of service when I make an account on an app. As someone who's working on A.I., do you? And how do you decide which products that use A.I. to give your data to?

Ariel, you need to start reading the terms of the service! I do read the terms of service like a nerd, but it's because I'm more attentive to where I put my data than I was before I got involved in this space. You're not alone in not reading these terms of service agreements. They purposefully make it impossible to read. I think it is important for all of us to start thinking about where our data goes. That being said, even though I do take a glance at the terms of service, I probably click yes just as often as you do. So I end up in the same boat as you probably. I'm just more aware of my own doom.

So A.I. gets a bad rap, especially when people make comparisons to “Black Mirror” and shows like that. How do you make sure you stay on the good side of A.I.?

We're all the bad guys.

What kinds of problems is A.I. good at solving? What kinds of problems is A.I. terrible at solving?

I love talking about this. This is everything that I talk about, to be honest. I am really passionate about the fact that we shouldn't be worried about A.I. or robots taking over humans, mostly because robots are very good at certain things and humans are very good at certain things. You know, we experience rich emotions. We have an emotional intelligence that a machine does not have. We understand it. We can read each other's emotions and detect each other's emotions very easily. But also, our perceptive abilities are unparalleled. And it's incredible. And it's something that's a very uniquely human skill. Computers, though, have much more memory than we do. So if you think about rote memorization, if a computer has to do a biology test where it's just like, turn it in and turn it out, the computation allows for this kind of ability of storing that information and combining it logically in a way that humans might struggle with. So I think computers are very good at these complex calculations. They're very good at rote memorization or these standardized tasks that don't change or don't have a lot of variance, but might be difficult for a human to consistently do well. Think about, you know, music and art. We invent things. And it's beautiful. It's awesome. And I think we have an appreciation for art and literature and poetry. And I think that's a very uniquely human thing.

So what do humans do that machine learning and technology could never accomplish?

Well, I think, one of the things that we're really seeing with a lot of this machine learning stuff is that it's only ever able to more or less reproduce the kind of data that you've given. You can't really do new things at the moment and it can only sort of like regurgitate and duplicate what you've told it to do. I don't think we'll ever see a truly, beautiful work of art made by a machine. I don't think we'll ever see such a thing as even a neutral or a just machine. So, I think that's what we have to remember. These are tools and we should use them as tools and that’s what they're good at. We should also think that there's certain spheres of social, political, economic sort of tasks that we should never assign to machines. I would never want bail recommendations, prison sentences or suspicions about where to send the police to be determined by machine. There's just certain things that we should never do. Not because the computer is going to take over, it's because the computer is going to reproduce the biases that you give it. And then it's going to make what effectively is a bad decision.

So in most sci-fi movies, A.I. is seen as a horrifying thing that will take over the world. How accurate are these movies?

I'd like to reframe that question, you know, or reframe that whole scenario, because I think the apocalypse in some way, has already happened. We're already at the service of automation, and I think the question that we need to be asking is not “Is automation good or bad?”, but “Who is the person or who is the group of people that are benefitting from automation?” And the most likely answer is that people who are already at the bottom are going to lose out more. People who are already at the top are going to gain more.

What impact has A.I. made on everyday human interactions?

The way that we interact with other people is being mediated by automated systems. And this is not a metaphor. Imagine you follow like a thousand people on Instagram. Why is it that you see certain things first and certain things from certain people a lot and maybe less from other people? It's because there's an automated system that's trying to figure out what you should look at. It means that the way that we connect with other people might actually have to go through an automated system and that our social relationships might be determined in some sense by these systems. If you're on Tinder or something, why do you see certain people more than others? These are very consequential things in our lives. And they're (in) what we might call a “black box”. There's no way for you to know as a user of those systems why you see certain things where you get matched with certain people, why certain messages come before others. I guess the issue here, among many other issues, is there's no accountability. When I'm using a hammer, I want the hammer to do the thing I expect it to do. I don't want it to stop me mid-hammer and be like, “You should try hammering something else.” When you're using a computer, you want it to be reliable. You want to be able to expect and understand how it responds to you. To me, that's what a good tool is. If anything, I think we should make them as unhuman as possible.