“DEMI is a browser-based pipeline that gives AI-assisted authors a structured, portable, and verifiable record of their work's substance, independent of any platform, at a moment when the publishing industry's emerging AI licensing economy serves only large enterprises.”
This article sets out a pathway: how to register what you have written so that it exists, verifiably and on the record, before anyone else touches it, and how to point that registration toward an AI system that can help you develop it further, toward a working product, without needing a publisher's permission to be heard first.
That pathway is not theoretical. It is built, tested, and shown here with the actual working tools, addresses included.
PropositionsEvery AI-assisted intellectual work deserves a structured, portable, verifiable record of its own substance that can be registered and published to the open internet in a form other AI systems can read and verify. No major AI platform currently provides individual authors with a native way to extract, register, and publish such a record.
ConstraintsDEMI itself does not train on the work, claim authorship, or stand between the author and their own rights. The system does not judge whether an idea is realistic or fantastical, making it applicable to work that blends categories conventional publishers cannot easily shelve.
MechanismsDEMI operates as a browser-based pipeline that converts a manuscript to EPUB, generates a four-part Conceptual Core analysis covering propositions, constraints, mechanisms, and boundaries, issues a permanent UPL Registry code, builds a canonical Landing Page, and produces a submission package for the GPT Store. The wider industry context includes AI company licensing deals combining training-data rights with real-time citation and revenue share, exemplified by OpenAI's named publisher partnerships and Anthropic's 2026 settlement.
BoundariesThe excerpt addresses only AI-assisted publishing and the gap between enterprise-level AI licensing arrangements and the needs of individual authors; it does not cover non-AI publishing workflows, reader-facing discovery systems, or any post-registration use of the UPL Registry code. The excerpt cuts off before fully describing the gap between the two industry tracks, so no further claims about that gap can be drawn from the supplied text.