The Library Of Congress Is A Training Data Playground For AI Companies

With archives hosting about 180 million works, the world’s largest library is drawing interest from AI startups looking to train their large language models on content that won’t get them sued.