What are “GANs (Generative Adversarial Networks)”

A simple explanation of GANs and how they work.

GANs, or Generative Adversarial Networks, are a type of AI model designed to create new content; images, video, or audio.

They work by pairing two systems and setting them against each other.

One generates content. The other tries to judge it. The generator aims to create something realistic, while the evaluator tries to work out whether it’s real or artificial. Over time, both improve through that back-and-forth.

That’s what allows GANs to produce things like realistic faces, artwork, and other synthetic media.

The interesting part isn’t just what they create, but how they learn. Instead of being told exactly what to do, the system improves by competing with itself.

Which raises a slightly different question: if something looks real, what actually makes it real?