Hello everyone, my name is Dora Agostino, and welcome to FinTech TV.
Today I have the honor of welcoming Doctor W.
Scott Scarnetta, a physicist, a pioneer, and one of the most important figures in the entire history of digital finance.
Back in 1991, Doctor Shornetta and his research partner Stuart Haber published a paper called How to Time Stamp a Digital Document.
That paper became the foundation of what we now call blockchain.
And here's something that tells you just how important that work was when Satoshi Nakamoto wrote the original Bitcoin white paper.
Three of the eight citations referenced Doctor Cernetta and Doctor Haber's research.
Scott then co-founded Surety Technologies in 1994, which launched the world's very first commercial blockchain in 1995, a blockchain that is still running today.
Now, as founder and CEO of Shermer.
Digital, he's applying those same ideas to one of our most urgent problems of our generation, deep fakes.
In a world where AI can make you see and hear things that never actually happened, Shermark gives people verifiable digital identities, so you always know who you're really speaking with.
Thank you for joining us, Scott.
Uh, happy to be here, and it really is me that you're speaking with.
A few years ago, I started a YouTube channel dedicated to teaching financial literacy to kids.
Now thanks to the wonderful team at FinTech TV, I have the chance to interview inspiring business leaders and hear their thoughts on money, the future of finance, and advice for young people just starting their financial journey.
Let's begin.
Kids my age don't know a lot about AI, but surprisingly, lots of kids my age are concerned about deepfakes.
Can you explain briefly what AI deepfakes are?
Also, though kids my age are scared of these deepfakes, most don't know the severity of this problem.
Can you put that into perspective for kids like me?
Sure, so, um, it's a great question, and I think when we say deep fake.
We're immediately trying to say.
Things that are done with artificial intelligence that are deliberately designed to deceive.
And while that's always been uh a situation that's possible, what we have is an extreme magnification of the possibility.
That things that you look at and see and hear that seem so real and realistic can be entirely made up.
They have no basis in fact whatsoever, and yet they feel like they're very, very real.
That's, that's the concern.
Blockchain helps Bitcoin by assuring that your wallet is yours.
Blockchain helps Shermark assure that your video is yours.
How does Blockchain help do this?
I think we need to.
Take just one step back in talking about blockchain because.
It's using the same principles that Doctor Haber and I used to develop the early blockchain.
We've used very similar principles to develop a solution for digital identity.
And so, it's not so much that we take blockchain and simply apply it, like trying to fit a square peg in a round hole, but we take the same principles we use to to solve the immutable record problem to solve the unfakeable identity problem, and the key idea.
Is that the the big innovation of the blockchain was to remove the need to have a trusted third party, somebody that You couldn't question somebody that you just had to trust because you had to trust them.
And instead, what Doctor Haber and I figured out is there's a way to distribute that responsibility for trust among the whole community.
That supports the task, and so, you don't need to trust any one particular person, because collectively, we all contribute to trusting.
And it's the same solution in that sense that applies to digital identities.
It's not so much that you prove someone is a particular person, because you use biometrics, or because you have a government issued driver's license or passport.
It's that the entire community vouches for who the person is, and in that same way.
As the blockchain, we've distributed the trust throughout the community, and it just makes it a safer and more difficult to fake.
Um, situation.
If I post a photo of myself and say I'm in Dubai on a trip, and 95% of the comments say, no, you're not.
I saw you yesterday, you're lying, you can assume that I'm not actually in Dubai.
This is like how consensus mechanisms help people assure, because if you, if you're a fake person who's stealing someone else's identity, and 95% of people say this is not really you, you're not who you say you are, then you can kind of assume that this person is lying.
I think it's a great way to emphasize the power of the community, OK.
And this is not unlike something that's available in various forms on different social media platforms.
The one I'm thinking of is Twitter, which has now become X.com, and they have something called Community notes, where you try to get the community to offer context to statements that are made, including controversial statements, and it provides a little bit of balance into the system.
In a sense, the only thing that's different between what Shermark does.
And what community notes are trying to do.
Is that Schrmark makes sure that the comments themselves that you're reading, the context, the community notes are secure, and that they're authored by the people who claim to be, you know, who they say they are, as opposed to someone anonymous trying to tip the scale by injecting things that aren't true.
Um, from what I know, Shermark allows you to mark your videos.
This is sort of like sending a transaction over blockchain.
Nowadays, almost everything can be copied with AI.
Is it possible?
Is it possible for AI to fake the mark?
And if so, how can you tell?
No, it is not possible to fake the mark.
Now, many people confuse what Shermark is doing with something called a watermark, namely, you take the video.
And then you add something visible in the bottom right corner, let's say, that says this is authentic.
The problem is, anything that you can add on in that way.
Uh, AI can also remove in the same way, or attach to a different video.
What's very distinctive about Schermark is that you use a credential.
Something that you create yourself and which only you can use.
To attest to the authenticity of a video.
No other computer, no matter how powerful, no other AI program, no matter how powerful, can provide the same proof that that credential provides.
And so, in that sense, it's very different than a watermark, and very, very credible.
Something I've been thinking about is how kids think of AI versus adults.
Now, I wanted to hear your opinion on this, but do you think that kids are better at detecting what's AI and what isn't since they grew up with it?
Well, they do grow up with it, and I think it gives them one advantage, which is they feel comfortable making use of it.
OK, that's the good part.
Um, as to whether it makes it easier for them to detect fakes or not.
I'm not so sure, and part of the problem is things are changing so rapidly, even two years ago.
There were certain indications that you could use to detect whether something was a fake or not.
That problem became that much harder a year ago, as the quality of the fakes improved dramatically.
And we're going to reach a point where if we're not already there, where there's simply no discernible way to detect the difference between a deep fake and an actual interaction, just on the basis of what's in the video.
That's why with Schurmark, it's something called out of band.
That's a technical term that means the proof is not contained inside the video or the live stream.
The proof is separate and distinct, is outside the signal.
And from our perspective, it's the only way that we're going to have a long term solution to this problem.
It's funny how you mentioned how AI has changed so much because my last question is, what is the next step for Shermark as AI gets more complex and more powerful?
How will Shermark evolve to protect our generation?
OK, so this is also a great question, and I'll pay you later for setting it up.
Um, So many people think that the way we're going to prevent deep fakes is by using AI to counter it, so that when the deep fakes get better, we'll make the detection software get better, and then when the deepfakes get better than that, we'll make the detection software that much better.
Well, that is an infinite cat and mouse game, and there's no way to win that game.
And so, with Schurmark, by focusing on the foundational principles.
Of authenticity.
We're not trying to detect.
Deep fakes.
We're not even trying to defeat deep fakes.
We're trying to do something very different.
We're trying to make deep fakes irrelevant, so that it doesn't matter how good the deep fake is, it just doesn't matter, because either you have the credential or you don't, and there's no middle ground.
This has been such an interesting conversation.
Thank you so much, and I hope you and everyone watching has a great day.
OK, can I just add one more piece of advice since this is a children's financial network?
Start saving now.
