Suno v5.5 Lets You Clone Your Voice. But Who Owns the Output?

On March 26, 2026, Suno released version 5.5 with three features that shift the platform from a text-to-music generator to a personalized creative tool: Voices, Custom Models, and My Taste.

Voices is the headline feature. Pro and Premier subscribers can record or upload their singing voice, verify it through a spoken phrase, and use that vocal identity to generate new tracks. The voice is private to your account. Custom Models lets you train a personalized version of v5.5 on at least six of your original tracks. My Taste runs in the background, tracking your patterns to surface personalized suggestions.

Together, these features mean Suno-generated music can now sound like a specific person. That changes the rights question fundamentally.

Before you can use your voice in Suno v5.5, you must check a consent box. It is not optional — without it, the feature does not activate.

By checking it, you grant Suno permission to use your voice data to train, develop, fine-tune or otherwise improve their AI and machine learning models. Not just your personal songs. The platform's models, broadly.

The consent is not limited to improving your personal voice model. It extends to Suno's broader AI infrastructure. Once checked, your voice data enters a training pipeline with no documented mechanism to retrieve or delete it.

For creators who have built a career on their vocal identity, this consent decision is a provenance event. It should be documented the same way you would document any other rights decision: with a timestamp, a record of what was agreed to, and proof of what terms were in force when the agreement was made.

The human authorship question for AI-assisted music has always turned on what the human actually contributed. The U.S. Copyright Office requires human creative control. Suno v5.5 adds a new dimension: if you clone your voice and use it in a generated track, your vocal identity is present in the output, but the AI is performing it.

Three possible interpretations exist. First, voice as raw material: the cloned voice is an input like a sample, and the AI's use of it is generative rather than performative. Second, voice as performance: the creator's vocal identity carries enough expressive character that its presence constitutes human contribution. Third, voice as context: the voice clone influences timbre but does not constitute a creative contribution in the copyright sense.

No court or Copyright Office ruling has addressed voice cloning in AI music generation specifically. Creators who use Voices should document their vocal contribution independently.

Whether voice cloning strengthens or weakens a copyright claim, the documentation requirements are the same.

Suno v5.5 is not just a feature release. It is a change in the rights landscape. The consent checkbox alters the relationship between creator and platform. The voice cloning capability changes what human contribution means in a copyright context. The Custom Models feature means the AI's output is shaped by your prior work in ways that were not possible before.

Each of these is a provenance event — a moment where the conditions of creation changed. A track generated on Suno v5.4 and a track generated on Suno v5.5 were created under different technical and legal conditions, even if the same creator made both on the same day.

Provenance documentation should capture not just what you created, but the version of the tool that created it, the terms that governed it, and the consent decisions that enabled it.

These dates mark shifts in the conditions governing Suno-created music.