Comment And that too is training (Score 1) 24
The company notes that consumption of these tracks is still very low, "between 1-3% of the total streams," and 85% are flagged as fraudulent.
All of the tracks identified as 'fraudulent' - along with the AI-generated ones that aren't so identified - end up training the AI to create more human-sounding music. And newer human song writers and musicians are making it easier for AI. Much of modern popular music is objectively less complex and distinctive. This probably isn't helped by the number of cooks preparing the broth.
Thirty or forty years ago it was very uncommon to have more than three writers for a song, and one or two writers was the norm. Today, it's very common for a song to have six or more - sometimes many more - writing credits. Even when you exclude vanity credits, there are still too many writers for the average song.
So a lot of originality, quirkiness, personality, and other positive attributes are suppressed and averaged out of today's song writing. Not to mention the lack of commitment and investment - who's going to toil and sweat over getting a song just right when several other writers' visions are also in the mix?
BTW, I'm just expanding a bit on a video I saw recently. It's fair to say that I don't know a lot about music or the industry - but you can't really say that about Rick Beato.