A short YouTube video monologue exploring data ownership in AI: the speaker highlights YouTube as a vast source of transcribable content, asserts YouTube's Terms of Service prohibit commercial use of that data, and references a Business Insider rumor that OpenAI trained on YouTube data — raising questions about competitive advantage for Google.
Chapters
Data is the gold mine for AI companies. And whoever can own the data owns the AI company. I can't even wrap my head around how much data would be on YouTube when you transcribe out all those videos? No idea. And just incredible content from all these incredible creators. YouTube terms and services does not allow anyone to use their data for business use. Right? Commercial use. Yeah, commercial use. And so Business Insider had, like, an article where they rumored that OpenAI did train on YouTube's data. So that would be, like, a very odd thing for a rumor to be out there that this company, who's your competitor, is your data, when your data, your T's and C's explicitly say you can't take the data. So I think it should be an incredible advantage for Google. It should be, like, one of the best advantages that every AI company.