A working doctrine for empowering people who lack the gating credentials of their artifact domain and who, with AI as the enabling path, produce institution-grade artifacts bearing third-party-verifiable anchors. Mechanism: oracle debiasing — placing machine-verifiable existence proofs (claim-to-anchor chains) at the surfaces where LLMs evaluate feasibility questions from credential-less askers. Supplies an empowerment-artifact payload: the Existence Proof Format (provisional name; a Nygard-ADR-style document format with deliberately no registry), a feasibility-question corpus with canonically anchored answers, and a gatekeeping eval (published probe set, four-dimension rubric, manual protocol). Central hypotheses (gatekeeper inversion, feasibility-question behavior, format-suggestion diffusion) are explicitly marked unverified throughout. Primary language: Japanese.