Image, video, audio and content generation.
A Generative Adversarial Network (GAN) is a machine learning model made of two neural networks that compete against each other to generate realistic new data, such as images or text.