AI Music Deals Are Making Headlines. The Rights Risk for Brands Hasn’t Gone Away.

If Suno v6 was built with licensed music, isn’t the rights problem solved?

It is a fair question. Suno says its v6 family was developed with Warner Music Group, BMG and Believe. Music Business Worldwide reports that the models were trained from scratch on different data, including licensed partner material and user-preference signals. Suno’s v6 FAQ also says every earlier model has now been retired.

That is a meaningful change. Licensing now sits closer to the model-building process, with rightsholders participating rather than watching from the sidelines.

But licensed training and cleared output answer different questions.

Licensed Training Is Not Output Clearance

A training licence governs how source material may be used to build a system. Brand clearance asks what appears in one particular track, who may own those elements and how the finished music will be used. A platform can improve the first without eliminating the second.

Commercial Rights Are Not a Clean Bill of Health

Suno’s current Terms of Service make that distinction visible. Eligible paid users receive certain commercial-use rights in downloaded outputs, but Suno does not promise that using an output will not infringe a third party’s rights. It also notes that outputs may not be unique. Commercial permission from the platform is therefore not the same as a clean bill of health for every campaign.

Exclusivity is another issue. The US Copyright Office’s AI report says purely AI-generated material is not protected by copyright unless sufficient expressive elements are determined by a human author. A brand may be allowed to use a track yet still have less control over exclusivity than it expects.

What Brands Still Need to Clear

The practical risk rises when a generated track contains recognisable protected material. If it reproduces elements of an existing composition, use in advertising or video may require permission from the composition owner. If it incorporates protected material from an existing sound recording, recording-side permission may also be needed. The US Copyright Office’s music licensing guide describes these as synchronisation and master-use rights.

From Generation to a Governed Audio System

This matters strategically, too. Brands need an audio system that distinguishes open exploration from approved use. AI can help teams search quickly, test directions and create variations. Final execution should still be anchored in assets and rules the brand can identify, document and reuse.

That is where an AI-powered audio production platform such as Sonica becomes useful. Teams can adapt approved music, voices and sound assets into short cues, edits and channel-specific versions. Reusing assets can reduce repeated production work; teams still need to review each version against the brand direction and its intended use.

Key Learnings

• Licensed training data and output clearance are separate. Training licences govern model development; brands must still assess the content and intended use of each finished track.

• Suno v6 marks a meaningful upstream rights shift. The model family was developed with industry partners, trained on different data and replaced Suno’s earlier models.

• Commercial-use rights are not a non-infringement guarantee. Suno’s terms grant eligible paid users defined rights while declining to promise that an output is unique or free of third-party claims.

• Brands need track-level governance. Preserve production records, screen for recognisable material, check current licence terms across channels and territories, and obtain specialist clearance for high-exposure work.