Interesting, I see this as 'testing the fence' behavior exposing an exploitable flaw in how Spotify(and I imagine a lot of other content providers) organize incoming media.
If it's that easy to poison datasets any moderate capability threat actor could conduct a coordinated attack. With use of decent effort could ddoss media providers in a way they cannot quickly (if ever) recover.
Interesting, I see this as 'testing the fence' behavior exposing an exploitable flaw in how Spotify(and I imagine a lot of other content providers) organize incoming media.
If it's that easy to poison datasets any moderate capability threat actor could conduct a coordinated attack. With use of decent effort could ddoss media providers in a way they cannot quickly (if ever) recover.
Alt. title from browser tab is less clickbaity, but too long for HN without editorialising
> 'AI slop': Emily Portman and musicians on the mystery of fraudsters releasing songs in their name