ChatGPT and Generative AI Are Hits! Can Copyright Law Stop Them?

As is now well established, ChatGPT and other generative artificial intelligence platforms are "trained" on huge amounts of images from the vast world of the Internet. Sometimes some of this content may come from the public domain, and other times, it may be licensed. Unfortunately, it is now confirmed that most of the content through which generative AI is "trained" belongs to neither the former nor the latter category.

In this video essay offered by Bloomberg Law, the company illustrates and explains the problem by starting from the foundation of the situation. It divides, in fact, the volume of images involved into "inputs," all those visual subjects that are taken from the world wide web, often, without asking for any kind of permission, and "outputs," the results that artificial intelligences generate and then disseminate online.