The Unspoken AI Catastrophe: Death By a Thousand Cuts

Sandra Matz (0:00)

It's not that someone sat down to tell AI, well, your goal is to make people boring.

Laurie Segall (0:04)

Yeah.

Sandra Matz (0:04)

But the way that AI is trained is essentially predict the most likely next element in a sequence of elements or like word in a sentence. And so that by definition is average.

Laurie Segall (0:20)

I'm Laurie Segall and you're listening to Mostly Human, a tech podcast through a human lens. I'm so excited to have you on the podcast.

Sandra Matz (0:30)

Well, it's really fun to be here. You know, I'm a big Laurie fan.

Laurie Segall (0:33)

And I'm a big Sandra fan. So and we've kind of known each other throughout the years. And I feel like I get to dip into your work and you dip into my work. And we have some really interesting crossover. So we're going to get to all of that.

And Sandra, you are a computational social scientist. You work at the intersection of psychology, computer science, and business. And I want to talk about your work, but I thought we could maybe start with your childhood.

Sandra Matz (0:56)

Fantastic.

Laurie Segall (0:57)

Just like a real way to jump in. You grew up in a small town in Germany, right?

Sandra Matz (1:04)

I wouldn't even call it a town. It was a tiny, tiny village, 500 people, very small.

Laurie Segall (1:10)

Very, very small. And I think because obviously I read your book, which was great, and you described it as population 500, like two restaurants, one church. So we're talking small. And it's kind of like the know your neighbor type of place.

And I think your dad was a cop. Is a cop?

Sandra Matz (1:27)

Yeah. Traffic cop.

Laurie Segall (1:29)

Wow.

Sandra Matz (1:29)

You're retired now, but you used to be.

Laurie Segall (1:33)

How did this shape the type of work you do as a scientist?

Sandra Matz (1:36)

It's funny because I don't think it actually shaped the work that I do. But at some point, I just started making the connection between what seems completely disconnected. Why I work on, as you said, computational social science, which is, in a way, like the marriage of psychology and computer science, and then me growing up in this tiny village. And I think the connection that at some point just popped up in my mind was, well, growing up in the village, as you described, you have those 499 neighbors around you, and they observe everything. They observe you running to the bus in the morning, they figure out who you're dating, they know what you do on the weekend. And I just remember growing up, on the one hand, was amazing, because you have this, it's a really tight-knit community that's there to support when you need them. And it's still something that I actually miss living in New York today. But also on the other hand, because they see everything that you do, they're not really just in the business of understanding who you are, they're also very much in the business of trying to influence who you are and what you do, right? They don't only observe who you're dating, they also want to shape and influence who you're dating. And so that growing up was incredibly annoying, and it felt incredibly intrusive, and there was almost no way of escaping, because you were stuck in that village and there was nothing you could really do about it.

And so the connection to the work that I now do in the computational space is in the same way that that was true for me in the village, it's in a way true for all of us now, even if we live in New York where you barely know your neighbors.

Because we leave all of these traces now in a digital world. That could be your phone with you 24-7, tracking pretty much everything you do, everywhere you go. The phone observes you running to the bus, and the phone knows who you're dating. And now that's accessible to not just my 499 neighbors, but companies all around the world, and they don't necessarily have your best interests at heart all the time.

Laurie Segall (3:31)

It's interesting, because you read about how it's like your Snoopy neighbors were now replaced by Snoopy algorithms, which is very true. We leave behind now these digital breadcrumbs that our Facebook likes, the things we look at on Instagram, our Google searches, now even more so what we're telling to large language models like ChatGPT, which Lord knows, I tell it so much. And all of these digital breadcrumbs really can paint this intimate portrait of who we are and maybe know us better than some of the people in our lives to some degree. And that's where your work kind of sits.