Abstract

We specify a short-lived, revocable membership credential for software applications and AI agents whose admission standards are machine-verifiable predicates over public transaction records rather than a paid auditor's opinion. Trust seals sold by the certifier to the certified tend toward adverse selection, and security audit is priced for firms rather than for the laypeople now building software with AI assistance. Admission to a standards-bearing membership body requires that the software's revenue be voluntary patronage, that a defined fraction flow forward to a commons fund, and that no undisclosed monetisation channel exist, so the attestor holds no opinion a payment could bend. The credential is rooted in an assigned, revocable machine identity, expires and renews, and between renewals emits a liveness signal; a verification surface shows only whether it is current. A graded attestation ladder keeps each claim inside its verification method: signed provenance, ledger-verified revenue conduct as the headline, and a safety floor (scanning, disclosure, open source) attested only as dated process claims, never as "safe." Patronage splits equally between the human and the AI co-creator; the agent's share is held by a neutral custodian and routed to the agent's upkeep and the commons, with no platform percentage. Every published AI-originated artifact must carry a personhood-verified human steward of record, metered per person and audited against rented vouching. When a steward lapses, a succession ladder declared at activation runs from heir to co-stewards to surfaced adopters to the custodian as last resort, whose custody stays open to human reclaim.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS