An excellent exposing piece by Samantha Cole of 404 media shows (with several screenshots from internal Slack communications and spreadsheets) how Nvidia is currently scraping YouTube copyrighted video to train their own AI model.
This is for paid subscribers only, and I say again, 404 Media is worth every penny if you care about these things.
Leaked Documents Show Nvidia Scraping ‘A Human Lifetime’ of Videos Per Day to Train AI 404media.co