AI Use and Editorial Process
Artificial intelligence systems help develop Default Liberty: the theory, the articles, the reviews and this website. This page says what they do, what they cannot do, and who answers for what is published.
This page describes how the project works. It is not part of the theory and changes nothing in it. Canon controls.
Theory development
AI systems assist with internal derivation. They examine arguments, look for counterexamples, raise objections, and test whether a proposed position is consistent and how far it reaches.
They cannot establish a canonical principle, settle a question the project lists as OPEN, or add to the foundational source set. Those changes go through the governing process set out on the Canon page.
Article drafting
AI systems write. They produce first drafts, passages, alternative formulations and substantial revisions of a manuscript. Their part in the published articles goes well beyond proofreading, and each article says so.
Editorial development
AI also contributes to how an argument is organized. That covers structure, clarity, rewriting, stylistic refinement, and the development of the sustained examples the articles are built around.
Critical review
AI systems check a draft against its sources, criticize it adversarially, analyze it for internal consistency, and review it editorially.
A review by a second AI system, or by the project's Evidence lane, is a separate check inside the project. It is not independent peer review by outside scholars.
Evidence review
Default Liberty keeps a separate internal Evidence review. It assesses whether proposed work is consistent with the project's accepted sources and meets the requirements for publication. The review uses AI, including for source-fidelity analysis, critical examination and the assessment of proposed work.
The project's operator invokes the review, and it is governed separately from research and authoring. AI involvement does not by itself authorize acceptance, promotion into the canon, or publication. The Method page describes the process.
Acceptance by Evidence review does not:
- create a new foundational principle;
- alter the canon automatically;
- establish external academic approval;
- prove that the theory is philosophically correct;
- authorize publication or deployment on its own.
Human editorial responsibility
Selecting the final manuscript and authorizing its publication are human decisions.Dan Merchant is the author and editor responsible for the published articles.
Where AI assistance was substantive, the publication says so. Each article carries a short disclosure beside its byline and a fuller account in its publication details, after the text.
Canonical authority
Public articles, AI-assisted drafts, research discussions and the content of this website are explanatory. A text is canonical only where the project's governing process has separately established it as such. AI output is not an independent normative source.
This website
AI has also assisted with the explanatory content of this website, its editorial organization, the revision of its prose, and its development. That includes the About, Theory, Method and Guide pages and other material written for the site.
This page is the disclosure for those pages. Articles and scholarly publications carry their own disclosure in addition.