the_viola•9d agoMost Upvoted

Just found out AI can spot fake luxury bags better than humans

I was reading this article from MIT Technology Review last night about a new AI system trained on leather texture and stitching patterns. They tested it on 500 bags from brands like Gucci and Louis Vuitton and it caught fakes with 97% accuracy. That amazed me because even experts sometimes get fooled by high-end counterfeits. Has anyone else seen this kind of tech being used for authentication in other areas?

3 comments

3 Comments

henryt189d ago

Got a buddy who collects vintage watches, nothing too fancy but he knows his stuff. Last year he bought what he thought was a legit vintage Omega from the 60s. Everything looked right to him and the seller had good reviews. He ran it through one of those AI authentication apps on a whim, just for fun. Turned out the crown and the movement inside were from a different era, something about the serial number not matching up with the model year. He was mad but he sent it back and got his money back. Saved him a couple grand for sure.

young.thomas9d ago

Oh man that is wild. I remember reading something about AI being used to spot fake sneakers and vintage watches too. Like they trained it on authentic Yeezy Boost 350s and Rolex Submariners, and it caught stuff that even seasoned collectors missed. I guess the AI can pick up on tiny differences in materials and construction that human eyes just can't see. It makes sense though, if you think about how consistent manufacturing is on legit stuff versus fakes.

thomas_price9d agoMost Upvoted

Hey hold on, I think there's a little mix up there. The AI wasn't trained on Yeezy Boost 350s for the sneaker part... that was actually a different project. I remember reading that they used data from authentic Air Jordans and some older Nike models to teach the algorithm what real stitching and sole patterns looked like. The Yeezy thing was more about verifying the boost sole material itself under a microscope or something. Still, your point about catching tiny differences in materials stands... that part is totally right.