Who owns AI-generated music?
It is complicated, evolving, and depends on the tool, the inputs, and the use. Read the licence, get permissions, and disclose.
AI music sits on shaky legal ground. The short version: in most jurisdictions, including India, copyright requires human authorship, which means a fully AI-generated track is hard to register and easy to dispute. Tools that allow training on copyrighted catalogues raise an additional layer of risk.
A practical posture for a working creator:
Use tools that disclose their training and licensing. Tools with clear commercial-use terms cost more but save you litigation risk and reputational damage.
For any voice work, get explicit, written consent before you clone. A voice is identity. Indian privacy and personality-rights jurisprudence is catching up quickly and being cautious is cheap.
For music in client deliverables, disclose. Many clients are fine with AI assistance — they want to know. Surprising them later is the failure mode.
For your own catalogue, treat AI as a sketching tool. Generate, then re-record, re-mix, and re-arrange with human performance and judgement. The result is more defensible commercially and creatively.
The field will keep moving. Disclosure, consent, and careful documentation will keep moving with it. None of that is exciting. All of it is what separates a professional from an enthusiast.