Here’s Proof You Can Train an AI Model Without Slurping Copyrighted Content

OpenAI claimed it's impossible" to build good AI models without using copyrighted data. An ethically created" large language model and a giant AI dataset of public domain text suggest otherwise.