Deepfakes are Coming for Your Job

Vijay Balasubramaniyan (0:00)

That one video, once we identified it, we were like, we're staring at the face of the new threat, and we need to start building software to be able to go after that.

Laurie Segall (0:17)

I'm Laurie Segall, and you're listening to Mostly Human, a tech podcast through a human lens. Welcome to Mostly Human. I'm so excited to sit here with you.

Vijay Balasubramaniyan (0:28)

Thanks so much for having me, Laurie.

Laurie Segall (0:30)

Yeah, and you know what's cool is we're in, just to set the scene, we're in Atlanta, which is where I'm from, but which is where your company Pindrop is located and where you live.

Vijay Balasubramaniyan (0:38)

Yeah, you know, it's funny. I came here to do my PhD at Georgia Tech, and then I fell in love with the city and never left, right?

I love the city because it's got a great food scene, it's got a great music scene, but most importantly, it's the hub for Delta, and so I can travel anywhere around the world without, you know, in a non-stop fashion, so I love this place.

Laurie Segall (1:01)

I'm a Delta gal, but I married a United guy, so it's a sensitive topic.

Vijay Balasubramaniyan (1:05)

There's a lot of loyalty point fights that you guys have.

Laurie Segall (1:10)

So I want to talk about Pindrop, your company which is a very successful company when it comes to security and identifying synthetic materials, and we will get into it, but I would love to start with Vijay before this company existed.

Let's go back. You mentioned Georgia Tech. You're a student at Georgia Tech. You go to a conference in India, and I guess you woke up and you said, I got to buy a suit.

Vijay Balasubramaniyan (1:40)

Yeah.

Laurie Segall (1:40)

And so you go try to buy a suit, and that transaction changes the course of your life.

Vijay Balasubramaniyan (1:45)

Yeah.

Laurie Segall (1:46)

Please explain.

Vijay Balasubramaniyan (1:47)

Yeah. So it's super interesting. I was actually getting published at security conferences here, and typically when you go to these conferences, you wear a suit, and as a grad student, I couldn't afford a suit here. So obviously when I went back to India, and India is where all these suits are made anyways, I bought the suit, gave it for alterations, and then I was flying out the next day. I get a call at 3 o'clock in the morning from my bank saying they want to verify a transaction. And I'm like, tell me the transaction. And they were like, oh, can you tell us your social security number so we can verify that it's really you?

And so we spent 30 minutes, and I'm trying to figure out if they're my bank, they're trying to figure out if I'm Vijay. And after 30 minutes, I said, go ahead and cancel the transaction. And then the next day, I'm packing, doing all of the stuff. And I call the suit company saying, hey, I want to come pick up the suit. And they're like, your credit card got declined, so we stopped working on it, and we can't work on it. So I came back to Atlanta really miserable, because the one thing that I went there for, I couldn't buy.

Laurie Segall (2:48)

You went all the way to India to get a suit.

Vijay Balasubramaniyan (2:50)

And you know, it's funny, when you're an immigrant, you have all of these things, weird thought processes going in your head, where things make you feel small, when they shouldn't make you feel small. And I'm like, I can't buy a suit, I can't afford it here. And so I was livid, and I was talking to my advisor saying, this is crazy that the phone system has existed for so long. And unless it's someone I really know, I don't know who's on the other end of that interaction.

So what do we do? And he said, why don't you make it the core of your PhD thesis? And that started our investigation. You know, you're an investigator. We started looking at what can we pick up in the audio that tells us something is truly coming from where it's coming. And that became a PhD thesis. It's now become a test of time winner, which in computer science is a really big deal. There are papers after 20 years that are demarcated as 20 year, but after 20 years, it's still a very relevant paper for computer science and has advanced the field. Pindrop is one of those papers. And then that became the basis for us to start a company.

Laurie Segall (4:00)

Wow. I mean, that's unbelievable that that and you hear this from so many founders, right? Like, it was just this problem. I don't understand why this is this exists or why this doesn't exist. And so you going back and forth and you guys just trying to say, prove you're human, prove you're real, prove this isn't fraud, was the beginning of what became not only just really important research, but a, you know, is it worth a billion yet? Are you worth a billion?