AI Music Tool Tracks Licensed Recordings Back to Artists
The Sonilo and Musical AI partnership is announced, not yet shipped, and aims to build attribution and payment into generative music from the start.
Updated on Nov. 3, 2026 in Artificial Intelligence
Live Poll
Do you think AI music companies should have to pay the artists whose work they use?
Sonilo, a San Francisco generative audio company, has partnered with Musical AI, a Canadian rights technology company, to track how licensed recordings contribute to AI-generated music. The partnership is announced; it is not yet a shipped product feature, and Sonilo is still finalizing master recording and publishing agreements with major rights organizations.
Why it matters
As AI-generated music becomes mainstream, attributing output back to the artists, songwriters and rights holders behind it is a hard technical problem that goes beyond simply securing permission to train on licensed material. The partnership reflects an industry push to build licensing, attribution and compensation into generative music models from the start rather than retrofitting them later.
The announcement cites no benchmarks, accuracy rates or performance metrics for the attribution technology. Sonilo's generation models are described only in general terms as video-native models built on licensed source material.
The players
Sonilo
A San Francisco-based generative audio company, backed by B Capital, that builds video-native music and sound models trained on licensed source material.
Musical AI
A Canadian rights technology company specializing in AI music attribution and rights management, which measures how licensed recordings contribute to generated outputs.
B Capital
A venture capital firm and backer of Sonilo.
The details
Sonilo builds generative models that create music and sound effects from video footage, text, or both, and says its models are built around licensed source material. Musical AI's technology analyzes each generated output and estimates how much the licensed recordings and compositions in the training material contributed to it, a process known as attribution. That measurement gives Sonilo's licensing partners usage data they can use to account for rights holders and support compensation.
Timeline
November 3, 2026: Sonilo and Musical AI announced the partnership.
Pending: Sonilo is finalizing master recording and publishing agreements with major rights organizations.
Pending: Sonilo is preparing to launch an update to its music model.
The Tech Race
Attribution for generative audio is an early-stage problem, and this partnership positions licensing-first music generation as an alternative to models trained without clear rights arrangements. The source does not name competing attribution efforts, so the broader race is described only in general terms.
Musicians, songwriters and other rights holders whose recordings are licensed for AI training may eventually receive compensation tied to how their work contributes to generated output. The source gives no pricing, availability or timeline for end users to access the attribution system.
The takeaway
Watch whether Sonilo completes its master recording and publishing agreements and ships its music model update, since those will show whether licensing and attribution work at scale in a commercial product. Also watch whether Musical AI publishes measured accuracy figures for its attribution method.
What happens next
Sonilo is finalizing master recording and publishing agreements with major rights organizations and is preparing to launch an update to its music model, though no dates have been given for either.
Further reading
For more on the models behind tools like this, explore our Artificial Intelligence coverage.
Source note: This article includes information reported by Broadway World.
Live Poll
Do you think AI music companies should have to pay the artists whose work they use?
