- Where Music's Going
- Posts
- License Your Voice to AI
License Your Voice to AI
PLUS: LNRZ collective is a launchpad
Hey there, you're reading Where Music's Going.
If you want to join 3,728 artists & builders learning how to navigate music's new future, subscribe below.
Welcome, artists & builders.
I’m trying something different. Shorter emails, 2x this week. Today has a case study. The next will cover tools. Let me know what you think.
Inside Issue #8:
💡 Context: 6 steps to license your voice to AI
🦁 In The Wild: Collectives are the new labels
📊 Numbers: Human curation vs the algorithm
💡CONTEXT
6 Steps to License Your Voice to AI
Midjourney prompt: a diagram. licensing your voice to an AI model.
On Monday Grimes said she will split royalties 50/50 with any successful AI generated song using her voice. I can see this future. In fact, licensing your voice to AI is music collaboration at scale.
Here is a framework (with inspiration from Holly Herndon’s Holly+ model).
1. OPEN SOURCE YOURSELF
Choose how you want creators to tap into a model of your work:
Build models directly
License to a 3rd party platform
Provide API access to your stems
The short-term Grimes patch: let your rules be known & put the pieces together after. This is why we need licensed solutions.
2. DETERMINE RULES
Define clear rules on how creators can use your vocal likeness.
→ Do they need approval for commercial release?
→ Can they pair it with other media?
→ What topics are off-limits?
→ What are the profit splits?
→ How do they label it?
Build a framework that is zero trust & fair.
3. VERIFICATION
Artists need ways to machine-verify a song was created using their work. The blockchain is very useful in this & may be a long term solution. But most things are not on-chain. For now, incentivize 'the right way' & have clear metadata & distro standards.
4. ENFORCE SPLITS
Once again, blockchain is useful.
Holly+ requires approval for commercial release via Holly+ DAO with on-chain royalty splits.
Grimes is having creators register on her Elf Tech platform & they will collect & pay out royalties direct to anyone who uses A.I. Grimes.
For a scaleable solution, we may need open distribution methods where you select a work's originator & split royalties.
5. BRAND IT
Provide naming guidelines that make it obvious music is modeled on you, but isn't actually you.
Holly+
DrakeAI
A.I. Grimes
Use the power of exponential name distribution, while maintaining clarity on what's official. It also provides clean artist pages where everything lives. Here’s the Grimes page. Here’s the A.I. Grimes page.
6. REAP THE REWARDS
The more copied the artist, the more valuable the original. The Beatles are the most covered artist in the world. Every new derivative work further cements their legacy & makes them more evergreen.
You set the rules.
----------------------------------------
I understand the hesitance with generative AI. It's new. It's scary. But it's an opportunity.
Thousands of AI deepfakes are going back & forth in Discord servers right now. This is just the start. The ability for artists to license, control, and monetize this - even if it’s entirely on a non-commercial release level - is a viable option.
Most of these will be made for an audience of one. That interactivity can be a key to unlocking more revenue & fandom for artists.
— Rob
🦁 IN THE WILD
Collectives Are the New Record Labels
Artist ran. Development focused. Collector inclusive. Writing camps!
LNRZ is a launch pad for artists to build their careers in web3 & beyond.
How LNRZ works: Artists apply for inclusion in a ‘class’ or a curated weekly drop. Members pick artists, vote on release details, plus receive rewards & access. Core members provide music direction, production, visual art & operations.
For each release, 80% goes to the artists & 20% goes to LNRZ.
The Freshmen Class: LNRZ brought 5 community-picked artists to Joshua Tree for a 5-day writing camp with producer & LNRZ core member Clear Eyes (aka Jeremy Lloyd of Marian Hill). The result became…
SATELLITES: a 6-track EP released as a digital collectible with some interesting gamification & incentives. Members voted on rarity, # of editions & pricing. Then had discounted pre-sale access.
Editions: 1,250, each holding 1 track with different rarity (1% to 54%).
Reveal: 3 days after buying, it’s revealed which song yours holds.
Curate: The community shared unique links to spread the word. The top 10 on the leaderboard received a bonus track/collectible.
Remix Competition: Stems & samples are available via permissionless, on-chain sample platform Kits, with the winning remix becoming the following week’s curated LNRZ drop.
Collect SATELLITES & you’ve unlocked entry into LNRZ HQ. You are now a member.
The Results: Sold-out of 1,250 editions, $22k+ earned. A growing community.
🔥 This is real artist development & real human curation, aiming for real community interactivity. It’s early days, but LNRZ are onto something.
More collectives / communities to dive into:
📊 IN NUMBERS
54% of listeners want music chosen by humans, not algorithms.
At least some of us want a human touch, says a Midia Research survey of UK consumers. Also:
38% of music steamers struggle to find music.
64% of TikTok users rarely know which artists they’re listening to on TikTok.
What Did You Miss?
Record Music NFT Sale on Sound: Snoop Dogg’s ‘Golden Egg’ secondary sale of 20 ETH (~$38k) Wed night 👀
‘They Tried to Shut Us Down’: Viral AI Bad Bunny & Rihanna track from a copycat of Ghostwriter977 had everyone fooled 👻
Avenged Sevenfold Join Grimes: Letting producers use their ‘sound’ to create new songs 🤝
Colors Joins Web3: The incredible music session brand drops free, open-edition NFT “marking the beginning of the COLORSXSTUDIOS X ZORA onchain journey” 🔵
Spotify is Sunsetting Historical Song Data: Download your stats now 💾
I’ll leave you with..
Was this forwarded to you? Sign up here.
Reply here with tools, case studies or ideas.
How was this week's email? |