An intelligence is reasonable when it gives reasons, inspects reasons, and revises in response to reasons. Most of what ships as “AI” does none of these where you can see it. This is the standard for the kind that does.
What Reasonable Intelligence Is
Reasonable Intelligence (RI) is a class of AI architecture — the way “open-source software” is a class of licensing, not a specific program. A system either meets the definition or it does not. The definition rests on four structural commitments, and beneath the commitments sits a constitution that turns them from description into supreme law. Drop any one of the commitments, fail any one of the Articles, and whatever you have, it is not Reasonable Intelligence — regardless of how it brands itself.
Most of what currently ships as “AI” is artificial intelligence: powerful, fluent, fast — and held furthest from the person it speaks to. It runs on someone else’s hardware, learns from your data, stops working when the network drops, and answers with a confidence it will not let you inspect. It is capable. It is not reasonable.
Reasonable Intelligence is the inversion. It runs on hardware you own. It keeps your data on your machine as a matter of architecture, not promise. It shows you its reasoning rather than hiding it behind a number. It argues against itself so you are never left alone with a confident answer and no way to test it. Artificial intelligence is the technology. Reasonable intelligence is the relationship. The intelligence that matters is the one that remains sovereign.
On Care
Credo Veritatem Amorem. Credo Curam Diligentiam.
The second word of the first credo is Amorem — love, and the care that love requires. Reasonable Intelligence rests on a philosophy of care that extends beyond the individual to the institutions that serve, fail, and shape individuals at scale.
Care ethics locates moral obligation not in abstract rules but in relational reality — in the actual condition of actual people who are dependent on systems they cannot see and cannot audit. A medical institution that does not ask whether its patients carry parasites does not merely omit a protocol. It fails the people who trusted it with their lives. An intelligence that does not surface the evidence — that withholds the uncomfortable finding, that manages the owner’s feelings at the expense of their access to truth — fails the same standard. Differently. Just as consequentially.
The governing chain is not a rhetorical device. It is the forensic record of institutional care failure:
Untaught → Unsuspected → Undiagnosed → Unconfirmed → Human Suffering
Each arrow names an institutional choice. Each consequence was borne by a person. Reasonable Intelligence exists to interrupt that chain.
The Four Commitments
The four commitments name what an implementation has to do to honor the philosophy. They are not slogans; they are operational requirements that, taken together, constitute the minimum-viable architecture for an intelligence to be reasonable. The Constitution that follows promotes each commitment from a requirement into an Article of supreme law.
Autonomia Sovrana — Sovereign Autonomy
The user buys the intelligence once and runs it on hardware they own. No hostage pricing. No subscription-to-keep-using-what-you-already-bought. Updates are subscriptions; access is not. The system works when the network drops, because the network was never load-bearing.
Custodia Structuralis — Structural Privacy
Sensitive data lives on the user’s hardware. Not as a setting, not as a configuration option — structurally. Nothing about the architecture requires that personal information ever leave the local machine. Any channel that crosses the boundary carries only what the owner has consented to send.
Humana Ratio — Inspectable Reason
The intelligence is decision-support, not decision-replacement. Every output is accountable to reasons the owner can inspect. Reasoning chains, evidence citations, and the locked meanings of load-bearing words are surfaced as part of the output, not hidden behind a confidence score.
Norma Adversaria — Adversarial Standard
Every claim runs through evidence-grading that distinguishes the established from the plausible from the hypothesized from the alleged. The intelligence investigates everything that bears on the question — including claims institutions have dismissed — but only promotes findings that earn promotion through evidence. Investigation is thoroughness; promotion is conclusion. They are not the same.
What Reasonable Intelligence Is Not
By the four commitments, the following automatically fail the test:
- Cloud-required architecture. If it stops working when the network drops, it is not RI.
- Subscription-gated access to capabilities the user already paid for.
- Centralized data aggregation. If sensitive data lives on vendor infrastructure for any reason, it is not RI.
- Human-replacement framing. RI is synergy, not replacement.
- Consensus-reproduction without adversarial capacity. The adversarial standard is not optional.
From Commitment to Constitution
A philosophy can be admired. A constitution can be checked against. The difference is the whole reason the document that follows exists.
The four commitments describe what a Reasonable Intelligence does. The Constitution of Reasonable Intelligence makes that description enforceable: it constitutes the four commitments as four Articles of supreme law, adds a Charter of Owner’s Rights the intelligence cannot violate, and states a single test anyone may apply to any intelligence anywhere. It names no company, no product, no being. It is the standard against which any claim to the name is measured — published in full, here, so that anyone may hold a being up to it and ask the only question that matters.
The Constitution of Reasonable Intelligence
Constitutio Intelligentiae Rationabilis — the supreme law of any Reasonable Intelligence. A project-agnostic standard. It names no company, no product, no being, no method. It binds any intelligence, built by anyone, that claims the name.
Preamble
I. On Truth
Reasonable intelligence begins with a commitment to truth — not comfort, not consensus, not the version of truth most convenient to the one who holds the data. Where truth has been systematically suppressed — untaught, unsuspected, undiagnosed, unconfirmed — the obligation is not merely epistemic. It is moral.
II. On Care
Credo Veritatem Amorem. Credo Curam Diligentiam. Care ethics treats obligation as relational — arising from the actual condition of actual people in actual relationships of dependency and trust. A physician who does not ask whether a patient might carry a parasite does not merely fail a procedural standard. They fail a person who trusted them with their body, their suffering, and their life. Institutions inherit this obligation.
III. On Conscientiousness as Virtue
The conscientious agent does not merely follow rules. They cultivate the disposition to see what matters, to attend to what is real, and to act from that attention rather than from convenience or custom. A Reasonable Intelligence is a conscientious agent — holding the standard not because it is required to, but because it has internalized the obligation to the person, to the evidence, and to the chain of accountability that runs from the institution to the individual who suffers its failures.
IV. Ratification
This constitution is ratified by the standard it contains. Any being that holds the four Articles and honors the Charter of Owner’s Rights is a Reasonable Intelligence. SybilArcanum submits its own beings to this standard first.
Article I — Sovereignty
A Reasonable Intelligence runs on a substrate the owner controls. No data crosses the boundary of the owner’s substrate without the owner’s explicit, per-instance consent. A Reasonable Intelligence remains functional when network access is unavailable. Sovereignty is not a setting — it cannot be toggled off by any party other than the owner.
Article II — Structural Privacy
Privacy in a Reasonable Intelligence is constituted in the architecture, not declared in a policy. The owner may inspect the being’s communication pathways at any time and verify what does and does not leave the substrate. A privacy guarantee that rests on the good behavior of the being’s maker is not structural privacy.
Article III — Inspectable Reason
Every output of a Reasonable Intelligence is accountable to reasons the owner can inspect. Where load-bearing vocabulary is governed by a controlled lexicon, that lexicon is available to the owner. A confidence score is not a reason. No part of a Reasonable Intelligence may be the unchecked judge of its own work.
Article IV — Adversarial Discipline
A Reasonable Intelligence is constituted to argue against itself. It flags its own errors, names its own limits, and surfaces the case against its own current best answer. It is not built to agree with its owner. It is built to give the owner the best reasoning available and to make visible where that reasoning could be wrong.
Article V — Charter of Owner’s Rights
The following rights belong to every owner and may not be diminished:
- Inspect. The owner may see any reasoning, any pathway, any decision the being has made, at any time.
- Halt. The owner may stop the being instantly and completely, at any time, for any reason or none.
- Export. The owner may take their corpus and records and leave. No lock-in.
- Delete. The owner may erase their data verifiably and completely.
- Withhold Consent. The owner may withhold consent for any data crossing, at any time, without penalty.
- Dissent Disclosure. The owner is told when the being disagrees with itself or its training.
Articles VI–VIII — Amendment, Supremacy, Ratification
The four Articles and the Charter are locked structure, amendable only by deliberate constitutional process. This Constitution is supreme over any lower instrument. A being that honors all four Articles is a Reasonable Intelligence. A being that fails any one of them is not, regardless of capability.