Rightsify LLC, a Pasadena company that gathers large datasets of music information and inputs it into computers so they can learn how to create music by themselves, has introduced its ‘Global Copyright Exchange,’ which it calls the world’s first AI music dataset licensing service.
The Global Copyright Exchange, or GCX, is a huge collection of music data that users can tap for many things, such as making their own songs, finding new music, or making music for games and videos.
Rightsify LLC said GCX has retained the services of trained musicologists to very carefully sort and label the music data so computers can learn from it easily.
“We’re excited to offer our data to companies launching innovative new music products and services, and with the AI and machine learning research community,” Alex Bestall, founder of Rightsify and GCX, said. “With the advent of large-scale music models, we believe our datasets will catalyze a revolution in accessible music creation, ushering in a new era of diverse musical applications and experiences.”
Large-scale music models are machine learning models that can generate, analyze, or understand music using large amounts of data. Rightsify said GCX has more than 32 billion parameters and tracks.
In practical terms, with such a huge amount of data, GCX can offer accurate and consistent annotations of the musical features and characteristics of each track, the company said.
A Rightsify LLC statement said GCX can power a variety of uses including DIY music production, music generation for ads, background music for videos, and adaptive music for gaming. It can also make music APIs (application programming interfaces) perform better such that music enthusiasts can easily search for music from all available online resources and obtain licenses to be able to use them in their production.
GCX’s introduction comes in the light of interesting news in the music industry, such as the launch of a Human Artistry Campaign in order to support musicians and songwriters in the face of growing AI capabilities, and the U.S. Copyright Office issuing decisions around AI-related works saying that the author has to be a human and cannot be completely computer-generated.
The Universal Music Group has also recently told music streaming platforms, including Spotify and Apple, to block AI systems from scraping its music; the company is in early discussions to license its songs to generative AI companies.
These steps are part of the process to maintain a balance between artists’ rights in protecting their work and creating something new, which is being challenged by AI-generated music.
Rightsify LLC says they’re making sure everyone can find a firm foundation for innovative and exciting AI music use cases.
“The era of large-scale music models demands an entirely new approach to datasets, especially in terms of copyright clarity and flexibility for refining groundbreaking models and apps,” Bestall said. “Legal and high-quality datasets should not be a stumbling block on the road to better generative music models.”
For more information about Global Copyright Exchange, visit www.gcx.co.