EP4-2 |Tired of Losing the Taylor Swift Ticket Battle? This Startup in Taiwan to Fight Scalpers

All right.

So I guess in this kind of startups, one of the key issues that could potentially arise is the privacy, which is a lot of people would say, oh, then this is this is probably another one of those scan your face technology.

And then the data probably sits somewhere inside.

And then my privacy is now, quote unquote, violated.

How can you verify that someone is, number one, is real and unique person without creating another database of these specific facial ID or facial photographs, facial imprints?

Yeah, yeah, sure.

That's a good question.

Just give you a quick background.

Actually, I started working on AI projects 26 years ago, and reviewing my age, but essentially a long time ago, with IBM, the research lab, Watson Research Lab, and also Toronto Lab.

So back then, we already started working on a lot of proprietary algorithm with text analysis, object recognition, facial recognition, So at the end of the day, facial recognition is not brand new, right? And it's been around, it's been used quite heavily in different parts of Asia, including China, and also a lot as well in the US and other countries, right? So the way we think about it is like, yeah, like you said, right? We want to protect our privacy, right? But we also want the convenience.

We also want the trust to create this unique humanist database with ticketing platform.

So the core of how we think about fan ID is that we don't want to simply solve identity problem by creating another new privacy problem.

When someone completes their facial verification, the system extracts your mathematical information of your facial features.

And then we turn them into digital representation.

So we never actually send your image or your photo to the server.

So the original image doesn't need to become a permanent identity database, just like you described, right? So the magic would be like, we use our sophisticated privacy preserving cryptography algorithm, including our serial knowledge architecture to protect the process.

So a lot of this processing is actually done on the user phone when they do the registration.

And so we can prove the person is the one holding the ticket, but we don't necessarily need to know a lot of personal information.

So I give you a, you know, basically a quick analogy, right? So a lot of facial recognition is a bit like a color symmetric matching, right? So assuming your face is red, and then most of the facial recognition tag or renders will detect the red color and send your face photo to the server.

So the next time when you do the face scan, they would match red with red.

Yeah, that's fair.

Obviously, there's a lot of still complexity involved.

There could be variation of red and stuff like that, but let's assume a simple analogy like that.

But that wouldn't have that privacy protection.

But in our case, we capture the red color and then our fan ID engine will do a sophisticated cryptographic processing locally on your phone and then conceptually is like mixing with another color, assuming it's yellow.