We can distribute some AI-assisted music, but only when it comes from AI services that are confirmed to use properly licensed training data. Other AI services are not currently eligible.
Here’s how it works:
• Our distribution partners require that any AI model used to generate the audio is trained on licensed, copyright-cleared material.
• Tracks made with AI services that have not confirmed this (for example, Suno, Udio, Mureka, and similar tools) cannot be delivered to any DSPs through our system at this time.
• The only AI music tools that are currently eligible are AI models that appear on the Fairly Trained “Licensed Model” certified list (fairlytrained.org).
If your track was created with one of these approved services, we can usually distribute it, but:
• It must be correctly tagged as “Includes AI” in the metadata.
• Some AI-generated content may not be eligible for certain UGC / rights-management platforms (like TikTok scanning, Facebook Rights Manager, or YouTube Content ID), even if it’s allowed on regular streaming services if it’s not considered unique enough.
What’s generally fine:
• Using AI to help write or brainstorm lyrics or musical ideas, as long as you own or clear all the rights.
• Writing your own lyrics and/or music (even with AI assistance) and then recording the final vocals and instruments yourself or with other human performers.
What we can’t distribute:
• Any recording where the underlying audio (music and/or vocals) was generated in whole or in part by an AI service that is not on the Fairly Trained certified list and has not clearly confirmed licensed training data.
We know this space is evolving quickly, and it can be confusing. We’re monitoring these rules closely, and if our partners update their AI policies, we’ll update our guidelines as well.