There’s a great little course from a few years back here, which you might be able to use for reference. There are also some useful tools with it.
https://www.kadenze.com/courses/machine-learning-for-musicians-and-artists/info
It focuses on relatively simple techniques which should have the advantage of being both accessible and immediate.
He’s a former OpenAI board member
I may be naive, but I honestly don’t think this was the original intention here. I think MS were quite happy having OpenAI at arms length as a maverick research group, while availing of the unlimited access to their results.
But it’s seemingly turned into more a case of Explode, Evacuate, Extinguish.
ChatGPT only exists because MS permit it. It’s not financially viable otherwise.
Yeah, they’ll keep OpenAI on life support for a few months until the exfiltrated staff have built MSGPT5, at which point Microsoft will just cut them off.
Yes, the recordings are 100% owned by the musician.
The songs themselves though, are not. If your focus is to generate new songs, then while there’s a good chance you’re in the clear, it is still an unanswered question. You’re essentially taking a bet that it’s legal.
If your focus is to be able to do a style transfer, i.e. turning music into something that sounds like it was performed by your pianist, then use of the data becomes something that is more easily covered under existing fair use.
I would suggest it may be something of a moot point though: can you actually build a useful musical model with this little data?