AI & Synthetic Media

GAN

Generative Adversarial Network — an AI architecture consisting of two neural networks that compete to generate realistic synthetic data.

What is a GAN?

A Generative Adversarial Network consists of two neural networks — a generator and a discriminator — that compete against each other. The generator creates synthetic data while the discriminator tries to distinguish real from fake.

GAN-Generated Images

Modern GANs like StyleGAN can generate photorealistic human faces that don't correspond to real people. These images often contain subtle artifacts detectable by forensic analysis.

Detection

GAN-generated images can be detected through frequency analysis, color inconsistencies, and characteristic artifacts in areas like teeth, ears, and backgrounds.