The surveillance scenario has a distinctive urgency: its infrastructure becomes invisible once normalized, so the objective is to set limits on inference and purpose before integrated agents make cross-domain inference routine. The projects below build on vocabulary the framework already holds — interior sovereignty, the right to opacity, and the rule that inferential access creates no jurisdiction — and turn it into machinery.
1. Surveillance jurisdiction map
Like the Gate Map, but broader: what is being sensed, by whom, at what resolution, for what declared purpose, what can be inferred rather than directly observed, how long the signal is retained, who else can obtain it, and which automated systems consume it. It matters because AI inference makes a traditional data inventory inadequate — knowing that someone collected location data says nothing about what they can infer from it.
2. Inference rights
Possibly one of the most important legal projects of the AI era. Current privacy thinking asks what did you collect?; the framework adds what are you permitted to infer from what you legitimately possess? A company may legitimately know what you purchased; that grants no standing to infer pregnancy, religious affiliation, financial distress, depression, political orientation, or the likelihood of a divorce. Inference itself can cross jurisdictional boundaries, and the capacity to infer is not a license to act on the inference.
3. Protected opacity zones
Deliberately preserve spaces where surveillance is technically or legally constrained — homes, libraries, some medical interactions, religious practice, political association, intimate relationships, some educational exploration, perhaps certain public spaces. The exact list matters less than the principle: a civilization needs zones where human activity is not continuously converted into institutional signal. These are not blind spots in a defective system; they are ecological necessities, because emergence requires somewhere to be provisional.
4. Anti-persistence architecture
One of surveillance’s biggest dangers is memory. Humans forget; institutions historically lost records; contexts disappeared. AI can eliminate that natural decay, so childhood behavior can follow someone at forty and every experiment becomes permanent. The framework resists this with deliberate mechanisms for expiration, deletion, non-retention, ephemeral identity, local processing, and aggregation rather than individual storage — because a participant must be able to become newly legible.
5. Ban or tightly constrain universal behavioral dossiers
The surveillance parallel to the universal participant score. The state knows one thing about you, the physician another, an employer another, your friends much more; an AI that integrates all of them creates something qualitatively different — the shadow participant, an externalized model institutions may come to prefer over the actual person because it is easier to deal with. That preference would be a major failure of the framework, and it is what makes this scenario the upstream source of gatekeeping.
6. Participant-side visibility tools
As with participant-side representation for gatekeeping, people should be able to ask: who holds information about me, what was inferred, which institution accessed it, which automated decision used it, where it propagated, and when it expires. This is a high-value use of personal AI, and it enforces the sixth protection directly — the field’s legibility to the participant must rise alongside the participant’s legibility to the field.
7. Surveillance firebreaks
Before the end of the decade, data must not flow frictionlessly between commercial systems, employers, insurers, banks, schools, governments, and law enforcement. Legitimate exceptions will exist, but the default cannot become if it exists, someone can eventually query it, because that turns every local observation into a potentially universal one — the mechanism by which the shadow participant is assembled.
8. Protect anonymous and pseudonymous participation
This becomes more important, not less. Reading, discussion, artistic exploration, political organization, and some inquiry do not all require persistent identity. The assumption that good people have no reason to hide who they are is incompatible with the framework, because opacity can be necessary for emergence, dissent, and ontological correction.
The timeline
Because surveillance normalizes invisibly, this scenario is at least as urgent as gatekeeping. The years will move with adoption; the sequencing is the point.
| Period | Transition objective |
|---|---|
| 2026–27 | Establish limits on inference and purpose — bounded observability, inference jurisdiction, retention limits, protected opacity, prohibited data combinations — before integrated agents make cross-domain inference routine. |
| 2027–29 | Build the technical firebreaks — local processing, compartmentalized identities, query auditing, data expiration, participant visibility tools, strict cross-domain access controls — so safeguards are architecture, not policy documents. |
| 2029–32 | Prevent ambient normalization — as cars, homes, wearables, workplaces, agents, and public spaces all sense by default, preserve genuinely unobserved pathways through ordinary life before opting out becomes impractical. |
| After ambient integration | Correction becomes much harder, because society has built safety, commercial, medical, and administrative dependencies on continuous sensing — and every reduction can then be framed as making society less safe. |
Guarding the safeguards
Each safeguard here can invert into the thing it opposes. A national data-rights portal could become the best map of everyone’s data. A universal identity-protection system could create universal identity. Audit requirements could force greater retention. A personal AI assistant could become the most intimate surveillance device ever built. So the safeguard is judged by the same standard as the thing it guards against, with no exception for benevolent intent. And the whole program is a set of requirements, not mechanisms: the framework can say a relationship must not acquire jurisdiction beyond its purpose, and it depends on encryption, secure hardware, differential privacy, cybersecurity, identity protocols, access controls, and constitutional law to make that real — machinery that becomes part of the finished civilization but is not derivable from the framework alone.