From there, we use AI to generate hundreds of possible versions of the idea: variations, mutations, accidents, echoes, voices, and alternate bodies. We are not asking the machine to decide for us. We are using it to expand the field.
When there are vocals, we may use AI-generated voices, but we write our own lyrics. The words are ours: our jokes, our obsessions, our arguments, our images, our taste. The voice is an instrument. The authorship is not outsourced.
Would we love real singers on these tracks? Absolutely. We want human performers. We want collaborators. We want voices with bodies, histories, flaws, instincts, and lungs. But right now, we do not have the money, the network, or reliable access to singers who actually show up. So we use the tools available to us instead of waiting around for permission, funding, or the perfect collaborator.