Institutional credentials function as pre-paid trust: a reader extends a degreed, affiliated author the benefit of the doubt because a recognised body has already vouched for them. People without those credentials have had no comparable way to earn a hearing, and their work has largely gone unread regardless of merit. This paper argues that generative AI changes that situation in two directions at once. It is widely told as a story of replacement — machines displacing knowledge workers — but it is also, less visibly, a story of recruitment: the same tools let people who could never have produced formal work begin to produce it. The question that follows is how such work can be trusted when its author carries no credential. The answer proposed here is earned trust through verifiable disclosure: where a credential buys trust in advance, transparency can earn it afterward, by making exactly how a work was produced open to inspection. The paper offers a concrete instrument for this — the AIast disclosure, a compact, controlled-vocabulary record of which AI systems were used, what each did, and how the work was checked, backed by producible evidence — and an optional, field-relative Lost Innovator (LI) prefix for authors working outside the credentialing system. It situates the instrument against existing practice (the CRediT taxonomy; publisher AI-disclosure policies), explains why structured disclosure beats prose, gives the incentives for credentialed and uncredentialed authors alike, catalogues the ways the standard can be abused, and states its limits. The central claim is modest in form and large in consequence: trust need not always be borrowed from an institution; it can be earned by showing the work. The paper is itself disclosed under the standard it proposes.
William Stafford (Tue,) studied this question.