Did Suno just use Blink-182 or Green Day tracks as training data to generate a generic pop punk ditty from California? Can anyone really figure out exactly which human songs, and how much of each, went into an AI song? Our guest Matt Adell, COO of Musical.ai (and former head of Beatport), says, yes they can. Their program determines who should get paid and who ought to get attribution when robots write songs based on human music. He shares with music attorneys Bill Hochberg and Sami Kazi where this new and interesting tech came from and where AI training is going in the music industry on this episode of Shmoozic Biz.

Table Of Contents:
0:00 Intro: the 12% Blink-182 Solution
0:58 Meet Matt Adell: Music Tech Legend
1:58 Matt’s History: From Napster to Beatport to Musical AI
5:05 Music Tech From Synthesizers to AI Music
10:29 The Jorja Smith AI Vocal Case
13:26 How Musical AI "Unscrambles the Egg"
15:00 Master vs. Composition Attribution Tech
18:05 Deterministic Similarity vs. "Magical Thinking"
26:10 Challenges in Detecting AI-Enhanced Tracks
33:40 Major Label Lawsuits vs. Indie Attribution
35:25 Why Licensing is Feasible, Not Inconvenient
35:00 Future Implications for the Music Industry

For news, views and shmooze you can use about the business and law of music, subscribe to Shmoozic Biz on YouTube and follow us on Instagram, Spotify (Video), and TikTok... links in the profile.

Bill Hochberg is a music and media lawyer at Raines Feldman Littrell LLP and he writes regularly in Forbes and other top publications on the business and law of music.

Our musical theme is "8 Church Birds" by billeeto. https://tinyurl.com/billeeto

DISCLAIMER: This program and the associated written materials are provided for informational purposes only and are not intended, and must not be taken, as legal advice on any particular set of facts or circumstances.