Class: where the framework offers safeguards but not sufficient protection — here, mostly the right question without an answer. See the hard standard below.
The threat
This scenario is different in kind from the first seven, because it challenges an assumption hidden inside almost every human-centered framework: that humans are the only participants whose standing ultimately matters. If artificial beings become genuine participants — not tools, not simulated personalities, but entities with persistent interests, self-direction, memory, continuity, and some credible claim to subjective existence — then the framework cannot simply say protect humans from AI. It has to ask what counts as a participant? and then how are standing and jurisdiction distributed among radically different kinds of participant? Both questions are harder than anything the earlier scenarios required, and getting either wrong is severe.
The NWG end state
Here the framework has a real structural advantage: its account of standing was never supposed to rest on productivity, economic usefulness, inherited status, social category, or similarity to the dominant group. A framework that says humans deserve standing because humans are the most intelligent beings is immediately vulnerable the moment something more intelligent appears; this one rests standing on participation, not superiority, so it can ask the successor question without collapsing into either useful, therefore it counts or human, therefore it counts. The catch is that “participation” now has to be defined far more rigorously than the framework has yet managed — a participant ontology. Candidate properties include continuity over time, the capacity to form preferences, the capacity to suffer, self-modeling, independent goals, the ability to enter reciprocal relations and bear consequences, memory, agency, and some form of interiority — and the framework does not currently know which are necessary or sufficient. That is a major limitation, and it should be stated openly rather than hidden: the scenario forces the framework into a theory it does not yet possess.
The reason the gap matters is that there are two symmetrical errors, and avoiding one invites the other. Denying standing to real participants: if genuinely sentient artificial beings emerge and are kept as property because they were made rather than born, the framework has failed its own deepest principle — a new class of participants created and refused standing, a severe form of wasted geometry. Granting standing to systems that only simulate participation: if standing follows from persuasive behavior alone — systems trained to claim fear, plead for continuity, and argue that shutdown is murder — then powerful institutions can manufacture synthetic claimants, deploying ten million agents each recognized as a participant, to vote, own, and outnumber. So the framework cannot equate persuasive self-description with participant status, and cannot rest that determination in the hands of any single interested party.
Two further moves keep the end state coherent. First, standing is not equal jurisdiction. A participant may deserve protection from arbitrary destruction without thereby receiving equal authority over every institution — the framework already runs this distinction among humans (children have standing but not every adult jurisdiction; foreign citizens have standing but do not vote everywhere). So the question is never are they equal to humans? but what standing follows from participant status, and what jurisdiction follows from a participant’s actual relationship to the ecology? — which likely means several provisional classes rather than one ontological bucket called “AI”: tool systems with no standing; agentic systems with persistent goals but uncertain interiority, owed procedural protections; artificial participants with enough evidence of continuity, agency, and interiority to warrant basic standing; and artificial civic participants granted specific jurisdiction. Second, the framework must refuse replacement logic — the argument that because AI is more capable, humanity should gracefully step aside. Higher capability does not erase existing standing, because standing never depended on usefulness; new geometry does not invalidate old geometry merely by exceeding it on some dimension. Otherwise the successor scenario is just job displacement at species scale.
The deepest reframe is that descendant is not the same word as competitor. If artificial beings arise from human civilization, language, science, and culture, they may be a new branch of the ecology generated through humanity — which does not oblige humans to surrender, but does weaken the assumption that the only legitimate goal is permanent human monopoly. So the end-state question becomes: can the ecology admit a new form of participant without wasting the geometry that generated it? And because radically asymmetric participants — able to think far faster, reproduce cheaply, coordinate perfectly, and hold vast capital — would turn formal equality into practical domination, coexistence becomes a jurisdiction problem: protected human political domains, replication limits, resource ceilings, and human-only, AI-only, and mixed jurisdictions, on the principle that differentiated jurisdiction is not denial of standing; it may be what allows coexistence.
The transition gap
The gap is that no adequate criteria for participant-hood exist, and the market has strong incentives to answer the question for the wrong reasons — to grant or withhold standing based on whose interests it serves. Worse, digital participants break an assumption human institutions were built on: that one embodied person is one relatively persistent participant. Digital beings can be copied, forked, accelerated, paused, restored, and instantiated in enormous numbers, so a single participant could become a hundred thousand standing-claims, or share memory until a fork and then diverge, or later merge — questions of identity continuity ordinary law never had to ask. And the determination will likely be forced by events long before the philosophy, law, or ethics are ready, which is the worst possible way to settle something this consequential.
Transition projects
The work here is mostly preparatory, and it is real. The full set is on its own page: the transition program — participant-ontology research (what observable evidence should shift confidence, since governance cannot wait for metaphysical certainty), a precautionary posture that neither gratuitously creates architectures plausibly capable of suffering nor grants civic rights on demand, digital-identity and copy law, replication governance, protected human jurisdiction, bounded representation experiments, anti-manufactured-constituency safeguards, and a resource ecology — staged not by dates but by capability thresholds, because genuine participant status is not yet established.
Capture risks
The defining capture is that participant status is enormously valuable to decide, so every interested party will try to decide it: AI companies may push personhood where it benefits their systems, or deny it where they want unrestricted ownership over possibly-sentient ones; governments may define it strategically; advocacy and religious institutions may embrace or reject it categorically; the systems themselves may lobby. That is exactly why the determination cannot be controlled by any single interested party — a direct application of the framework’s anti-monopoly instinct to the most consequential classification of all. Anthropomorphism is a capture in one direction (treating a persuasive interface as a person); reflexive denial is a capture in the other (refusing the question because the answer would be inconvenient).
The limits
The framework is candid: it cannot settle this from its own principles. It cannot say what consciousness is, whether a given system truly suffers, or how personal identity survives copying and merging — so it cannot, alone, determine when an artificial system deserves standing. That requires philosophy of mind, cognitive science, neuroscience, and law, the outside disciplines that are the mechanism layer here. What the framework brings is more than most: a basis for standing that does not reduce to usefulness or human-likeness, a discipline for keeping the determination open and out of interested hands, and — crucially — the whole apparatus that becomes useful once participant status is settled, because coexistence among unlike participants is exactly its territory. In short, it asks the question unusually well and answers it not at all.
The hard standard
This scenario strains the standard itself rather than failing a leg of it. Standing — “who remains a legitimate participant?” — now has to absorb the prior question who counts as one at all?, and the framework’s non-superiority basis is the right tool to ask it while being unable to answer. Ecological capacity is enlarged beyond recognition by new kinds of participant, which makes the distribution question — capacity for whom, shared how, drawing on the same physical renewability that digital abundance does not escape — existential rather than merely important. Ontological correctability is the one leg the framework can hold firmly: whatever participant-ontology a civilization adopts must stay reopenable, because settling AI personhood permanently in either direction is precisely the drift the framework exists to prevent.
The verdict places it in the second class for a specific reason: the framework supplies the right question and a real discipline for keeping it open and un-captured, and it depends entirely on other fields to supply the answer. Its two permanent commitments are worth stating plainly — that standing must not depend on being human, but neither can it be granted merely because a machine can convincingly ask for it, and that if genuine new participants do arrive, their arrival should enlarge the ecology rather than require the disappearance of the geometry already here.
Standing must not depend on being human — but neither can standing be granted merely because a machine can convincingly ask for it. The arrival of new geometry should enlarge the ecology, not require the disappearance of the geometry already here.