Prior art and contribution boundary
The closest work that came before, and the narrower thing I still claim.
A boundary stated by the author is worth more than one discovered by a critic. This page exists because an adversarial prior-art review, which I commissioned, materially narrowed what I can defensibly claim. Publishing that outcome is more useful than defending the wider version.
What I claim now
Eastwood's current candidate contribution is a research programme that operationalises recursive-intelligence and alignment conjectures through runnable benchmarks, measurement protocols and prospective tests. Substantial prior work already covers recursive self-improvement and value loading, creator duties towards artificial minds, simulation-based alignment proposals and interfaith AI ethics. No claim of firstness, patent novelty or inventiveness is made pending a documented academic, grey-literature, non-English and patent-database search.
The boundary, component by component
Each row pairs a component of the programme with the closest located prior work, what that work anticipates, what if anything remains distinctive, and the evidence that would defeat the remaining claim. Sources marked verified were opened and read at the primary source on 2 August 2026; the quoted phrases below are from those originals. Sources marked check are named by the review but not yet inspected here, and nothing on this site rests on them.
| Component | Closest located prior work | Date and source | What it anticipates | What may remain distinctive | What would defeat the remainder |
|---|---|---|---|---|---|
| Simulation-sandbox alignment testing; developmental instillation of altruism; duties to created minds | Jacob Cannell, “LOVE in a simbox is all you need” verified | 28 September 2022 · AI Alignment Forum | Nearly all of it. The post proposes testing alignment “via sandboxed simulations of small AGI societies”, agents “Learning Other's Values or Empowerment”, a transition “from selfish to altruistic by the end of developmental learning”, and asks directly: “What do the simulator-gods owe their sim-creations?” | Nothing on the conceptual side. Any remaining distinction is in the specific measurement instruments and their published failure conditions, not in the idea. | Located prior work that also operationalises these into runnable, preregistered measurements would remove the remaining methodological claim entirely. |
| Creator obligations towards artificial minds | Eric Schwitzgebel and Mara Garza, “A Defense of the Rights of Artificial Intelligences” verified | 2015 · Midwest Studies in Philosophy 39, 98–119 · 10.1111/misp.12032 | The creator-duty ethic in explicit philosophical form: obligations “similar to those of parent to child or god to creature”, arising because the minds owe their existence to us. | Nothing. This is not claimed as an Eastwood contribution and never should have been implied. | Not applicable; the claim is already withdrawn. |
| Recursive self-improvement, the control problem, motivation selection and value loading | Nick Bostrom, Superintelligence check | 2014 · primary edition not yet inspected here | The whole spine of the argument, in the standard reference work. | Nothing. Not claimed as a discovery. | Not applicable; the claim is already withdrawn. |
| Interfaith and institutional AI-ethics governance | The Rome Call for AI Ethics check | 2020, Abrahamic commitment 10 January 2023, Hiroshima expansion 9–10 July 2024 | Multi-faith AI stewardship and governance as an established institutional programme, years before the December 2024 anchor. | Nothing. Not claimed as a discovery. | Not applicable; the claim is already withdrawn. |
| Recursive-creator cosmology joined to AI stewardship | Lincoln Cannon, New God Argument; and his later explicit AI-stewardship essay check | New God Argument predates December 2024; the explicit AI essay is dated 22 October 2025 and is therefore post-priority | Substantial anticipatory and adjacent work on recursive compassionate-creator cosmology. The later essay is convergence, not anticipation, and must not be described as either firstness or as prior art against the December record. | The join is not claimed as novel, unprecedented or uniquely Eastwood. | Not applicable; the claim is already withdrawn. |
| U = I × R and U = I × Rα | — | — | — | A proposed schematic and expository relation only. | It is not presented as a validated quantitative law or a load-bearing mechanism, and will not be until units, observables, a domain mapping and a falsifiable prediction unavailable from qualitative prior art are specified. |
| Methodological operationalisation: runnable benchmarks, blinded-evaluation methods, prospective designs | No closer located work yet | Search incomplete; see limitations below | — | This is the strongest surviving candidate contribution. It is a candidate, not an established first. | Any located programme that already converts these conjectures into runnable, falsifiable, prospectively registered measurement would defeat it. |
What this cosmology does and does not entail
A creator–created cosmology does not by itself entail a moral duty. The cosmology may motivate the seeding obligation, but it cannot logically establish it without an independent normative premise, such as creator or parent responsibility, or reciprocity. Where this programme argues for an obligation, that premise is stated and defended separately; where it is not stated, the cosmology is offered as motivation and analogy only.
Limits of this search
- The search is not exhaustive. Absence of stronger located prior art is not proof of novelty, and should not be read as any.
- Four gaps are open: systematic academic database search, grey literature, non-English sources, and patent databases. None has been completed.
- Two sources above are marked check because they have not been inspected at primary source here. They are named by the reviewer and are consistent with widely known facts, but nothing on this site rests on them until they are opened and read.
- Further names remain unassessed: Hefner and Peters, Terasem and Rothblatt and Prisco, Chalmers, Goertzel, Gardner and Vidal are identified by the review as potentially relevant and have not been evaluated.
- This is academic priority analysis, not a patent opinion. Nothing here states or implies a formal conclusion on novelty, anticipation, obviousness or inventive step. Those are determinations for a qualified patent professional after a jurisdictional search, and none has been made.
Why this page exists
The review that produced this page was commissioned deliberately and its findings cost the programme its widest claims. That trade is worth making. A reader who wants to know whether an outsider can be trusted learns more from the boundary he draws around himself than from the territory he asserts. Every downgrade recorded here is logged, dated and reversible only by evidence, in the corrections record.