CaraComp
CARACOMP DAILY · EP.213

AI Deepfake Images: Seoul Official Fined · Video Briefing

A staff photo and free software. That's all it took.

September 11, 20263:10Watch on YouTube →
AI Deepfake Images: Seoul Official Fined · Video Briefing
Chapter 1 of 3 · AI DEEPFAKE IMAGES AT WORK
0:00 / 3:10

Full Transcript

A worker in Seoul grabbed his coworker's photo. It sat right there in the office directory. Then he used A.I. to fake it. He put her face on someone else's body. He made it look romantic. He posted it four times. He got caught, and fined six million won.

AI DEEPFAKE IMAGES AT WORK ▶ 0:14

A staff photo and free software. That's all it took.

This wasn't a hacker or a stranger. It was a colleague. He didn't break into anything. The photo was already there. That describes almost every workplace on earth.

According to SBS News, South Korean police caught four hundred nineteen deepfake suspects through July. These aren't celebrity scandals. They're happening inside real offices, made by real coworkers. And most bosses assume a photo is real until you prove it's fake.

So the victim gets hurt twice. Once by the fake, and again by having to prove it. A fake image doesn't need to be perfect. It just needs to make people pause.

The Seoul case ended in court. But most fakes never do, and proof is the hard part.

Picture your fourteen-year-old wanting a social account. The app asks for a selfie or her I.D. first. Now her face sits on a company server. You don't know who sees it, or how long it stays.


SOUTH AFRICA REJECTS AGE CHECKS ▶ 1:08

South Africa said no to mandatory age checks.

That's the fight South Africa just picked. The government looked at forcing age verification software on every kid under sixteen. And it walked away.

According to Biometric Update, officials found no real proof these bans keep kids offline. Minister Malatsi noted kids dodge the checks easily, often with a borrowed I.D. Meanwhile U.K. regulators fined one platform over fourteen million pounds, and the checks still failed.

So you collect a mountain of face scans and I.D. photos, for nothing proven. That's not a safety measure. That's a speed bump with a camera on it.

But if you can't inspect a photo, how does anyone prove an I.D. is truly real?

Look at the digital I.D. in your phone right now. It's really just a picture. A very official one. Someone could fake it. So why do banks and border agents trust it?


NATIONAL DIGITAL IDENTITY EXPLAINED ▶ 1:58

The proof is a signature you can't see.

The answer isn't how it looks. It's a hidden math signature buried in the data. It can't be forged, even with the file in hand.

According to Biometric Update, Zambia just built this trust system, called national digital identity. It uses two linked keys. A private key signs your I.D. A public key checks it. The public one can prove a signature is real, but it can never fake one.

Three African countries built this in four months. And here's the gift. It can prove you're over twenty-one without showing your name or address.

Three stories, one thread. A fake face fooled an office. A country refused to trust face scans it couldn't check. And Zambia built the math that proves a face is real. The whole fight is the same. Can you trust what you see?

Every source and today's podcast deep-dives are in the description. From CaraComp, see you tomorrow.

Sources

Stories in This Episode