# Ownership, Licensing & Provenance

**Track:** Go-to-Market in the Model Era — AI for Entrepreneurship — complete (29)
**Framework / surface:** venture strategy
**Level:** Advanced
**Prerequisites:** Moats, Wrappers & Commoditization

**In one line:** Inputs, outputs, weights, likeness, and feedback — the rights stack that determines what the business can actually sell.

## Theory, aesthetics & inspiration

Before the pricing page, do the rights audit: an AI-native venture is a stack of claims about ownership, and each layer has different law and different leverage. Outputs: the U.S. Copyright Office's guidance holds that copyright requires human authorship, so purely machine-generated work is not copyrightable — what your customers own in what your product makes is a contract question you must answer deliberately. Inputs: training-data rights, customer-data rights, and the increasingly valuable right to learn from feedback — the flywheel runs on permissions you must actually have. Terms above you: model providers' commercial-use terms and training opt-outs; weights below you: open-weight licenses that differ meaningfully in what commercial use they allow. Around all of it: likeness, voice, and publicity rights that creative products trip over first, and provenance — C2PA content credentials — that turns "we can prove where this came from" into both a compliance answer and a feature. The rights stack is not legal hygiene; it determines what the business can actually sell, and to whom.

**Founder question:** What, precisely, does your company own — and can you prove it?