To kick off the new year, we’re thrilled to launch W&M STARTER PACKS — a new series unpacking essential foundational concepts for navigating music and tech.
Every two weeks, we’ll highlight a timely topic in the music/tech news, and ground it in evergreen findings from our community-driven research projects. Hopefully, this format will give you more context on everything we’re doing as a community, and serve as a jumping-off point for further analysis as we collectively navigate the rapidly shifting music industry.
If you have any feedback or suggestions on this new format, please reach out to us at email@example.com — we’d love to hear from you.
Music and tech have had a dynamic, complex, and often fraught relationship throughout history.
On the one hand, music consumption, discovery, and fandom have been early use cases for virtually every consumer tech trend in the last few decades — from the first Internet forums and streaming services, to smart speakers, VR headsets, and the blockchain. The ubiquity and accessibility of music as a cultural art form make it particularly well-suited for scrappy experimentation with new technologies.
On the other hand, music rights are frequently positioned as major barriers to innovation. Every year, music-industry trade groups pursue nine-figure litigations against tech companies like Roblox or Peloton that build lucrative businesses off the backs of music catalogs without rights holders’ permission. The music startup graveyard is littered with countless examples of ventures that wanted to deliver on a delightful music fan experience, but ultimately burned most of their funding on content licensing fees. Music royalty collection also remains incredibly opaque and complicated to outsiders, adding a persistent layer of legal ambiguity on top of licensing costs that can be challenging for founders to navigate.
This tension will take center stage in 2023 as music-related use cases in Web3, generative AI, and the metaverse continue to gain steam. In all of these cases, developers are seeking open, accessible data — which doesn’t fit well with the permissioned, ringfenced approach that traditional music companies have taken to enforcing music copyrights online. At the same time, copyright is arguably the bedrock of how creators have been able to build brands and businesses off of their creative work, so the solution is not as simple as doing away with existing legal frameworks altogether.
In our research, we’ve delved deep into this turbulent relationship between music rights and new tech — outlining not only the key challenges that all stakeholders face (especially artists), but also some potential paths forward for solutions that effectively balance these competing interests involved.
The complex music royalty value chain
Before even introducing new tech into the picture, the music royalty value chain is incredibly convoluted and prone to errors.
There are two main reasons why paying artists and songwriters accurately and on time is still so hard:
- No central database for song ownership information. The reasons for this are both technical and political, including the reluctance of royalty collection societies to share and fact-check ownership data openly with each other. This lack of data standardization makes it hard for artists to ensure their rights metadata are reflected accurately across channels, and makes it hard for founders to know which rights holders with current authority to request permission from.
- More complex networks of collaboration and ownership. The average number of collaborators on a given Billboard Hot 100 song has increased over time. This has drawn out the song registration and subsequent royalty payout process; royalty collection societies require all composers’ and writers’ signatures to register a work, and these parties may be represented by several different rights organizations. In addition, more investment funds like Hipgnosis and Shamrock Capital are rapidly buying out music catalogs from the original owners, creating further accounting nightmares for rights holders and streaming platforms. Oftentimes, changes in ownership will not be reflected on the backend until months after a deal is closed.
Music rights and generative AI
The generative AI landscape is moving incredibly fast. In the last six months, AI tools like Stable Diffusion, Midjourney, and ChatGPT have collectively raked up millions of users and hundreds of millions of dollars in VC funding.
As music AI gains steam, we’ll have to revisit our entrenched notions of creativity, originality — and, by extension, copyright ownership. For the music industry navigating this new AI landscape, there are three main obstacles that lay ahead:
- No unified stance at the national policy level. The US is in the middle of potentially reversing a decision to grant a copyright to an AI-generated comic book, and the RIAA has flagged certain music AI tools as forms of piracy. In contrast, the UK allows AI to own copyrights, and recently introduced a policy to allow text and data mining “for any purpose,” with no opt-out for rights holders.
- No unified stance at the platform level. The lack of guidance from governments has left creative AI platforms to set ownership precedents on their own. Some, like Boomy, claim full ownership of all content generated using their tools, granting limited license to users. Others, like Midjourney, immediately grant paid users all rights to generated assets, while giving the platform a perpetual copyright license. Still others, like AIVA, only grant full copyright ownership to paid users on certain subscription tiers.
- Lack of clarity around training data for music AI models. While some music AI companies like Harmonai are transparent in how they train their AI models, including using CC-sign agreements directly with artists, other models like OpenAI’s Jukebox have been trained on large swaths of copyrighted songs without the original rights holders’ permission. This could potentially lead to accusations of copyright infringement for startups that choose to incorporate these models, and there remains no standard for how to share revenue generated from music AI tools with the people who provide the underlying training data.
Music rights and Web3
Across NFTs, social tokens, and DAOs, Web3 in all its forms has ushered a new wave of experimentation in the music industry, rethinking traditional approaches to music monetization and fan engagement.
According to our research, artists and music brands have generated over $200 million in primary sales from music NFTs in the last two years. This growth has also led to a string of lawsuits from both independent artists and major labels inevitably who are demanding their fair share.
That said, marrying copyright law to the blockchain is far from straightforward, for the following reasons:
- Ambiguity about what “owning” a music NFT actually means. Most buyers of music NFTs don’t actually own the copyright to the underlying music or artwork; rather, they are granted a separate, “ersatz” form of ownership established by the NFT’s smart contract. A lot of music NFT projects don’t make this distinction between digital property ownership and traditional IP ownership clear, which can lead to confusion and contradictions in legal agreement language — and, in turn, lawsuits.
- Limitations of the blockchain in reflecting the complexities of global music copyright. In the context of music copyright, NFT “smart contracts” are not actually contracts, and not actually that smart. While music NFTs effectively establish an on-chain relationship between the artist and the buyer, their underlying data structure is still behind in accommodating clauses that would be table stakes in traditional music agreements, such as territorial licenses and large-scale collaborative splits (especially splits that may change over time).
- Lack of metadata standards for music NFTs. One of the main benefits of engaging with Web3 is that there is a practically unrestrained design space for what information an artist wants to include with a given NFT. That said, this openness also leaves music NFT metadata fragmented and undefined; even simple fields like track titles, creator information, and media details that would drastically improve discoverability and indexing are missing from many music NFTs today. Developing a music NFT metadata schema from scratch will be a long, complex process requiring widespread consensus from multiple stakeholders, including artists, developers, and rights holders.
Music rights in the metaverse
The metaverse can be understood not as a singular thing, but rather as a series of desired states for how we interact online and offline — particularly facilitating more interactive social experiences, as well as more fluid movement and expression of individual identity. Several different technologies, including but not limited to 3D design, virtual and augmented reality, gaming, avatars, and livestreaming, can work together to make this vision come to life.
This presents an exciting playground for artists looking to build new immersive experiences around their music, foster closer relationships with superfans, and expand revenue opportunities for their work (especially outside of streaming and touring).
That said, several challenges persist in scaling music metaverse experiences in a way that brings all rights holders along for the ride:
- Complex licensing processes for live music metaverse experiences. Depending on the nature of a given music metaverse experience — is it a one-time in-game “concert”? an always-on virtual world? — artists and platform developers may need to obtain several different kinds of licenses to host the experience legally. These include but are not limited to master rights for any preexisting recordings included in the stream, sync licenses for rebroadcasting or making the stream available on-demand, and trademark licenses if any major brand names or visual imagery feature in the stream. It is not always clear who is responsible for obtaining these licenses.
- Unclear revenue models, especially for the long tail of artists. Partially due to ongoing legal battles, there are limited means for artists to upload their own music to metaverse platforms and access new monetization opportunities there. More lucrative placement opportunities like in-game concerts in Fortnite or Roblox are limited mostly to artists with existing major-label relationships — and even then, rights holders don’t get clear metrics or data tracking from metaverse platforms to determine whether their investments have any meaningful impact on fandom or consumption. As an alternative, many independent artists have turned to Web3-native virtual world platforms like Decentraland and Voxelsto bypass the traditional licensing system, instead monetizing fan engagement directly via NFT sales.