The music industry is building the tech to hunt down AI songs
VergeJUN 20
AI music, trained on artists' work without permission, endangers creativity and livelihoods, which is why the likes of Elton John, Billie Eilish, and Paul McCartney decry unauthorized use in LLMs. As thousands of lawsuits against AI startups expose rampant copyright violations, platforms must bolster algorithm safeguards, and regulators should mandate licensing to protect artists' rights and ensure fair compensation.
While the music industry should, and is, regulating AI music, it's also accepting the inevitability of this growing practice. With algorithms already being implemented to flag AI-generated songs before they go viral, coupled with licensing and compensation policies to ensure artists' rights aren't violated, this technology will become a boon to the industry without harming the humans at the heart of it.
This issue is bigger than just AI music; it deals with Big Tech's profit motives and faulty streaming analytics. The Velvet Sundown was able to game Spotify's "monthly listener" algorithm, which, rather than revealing genuine popularity, only shows how many times a song came across a user's page. Music corporations, backed by the government, aren't likely to change the system — whether it's gamed by AI or not — that benefits them.
There is a 66% chance that by 2050, at least 25% of Year-End Billboard Hot 100 songs will be primarily composed by AI, according to the Metaculus prediction community.