System map
The map describes recurring mechanisms—not a claim that every use is harmful or that one actor coordinates the whole system.
Removal is only one form of speech governance. Demotion, recommendation exclusion, labeling, reframing, and personalized invisibility can shape public knowledge without a visible decision.
Decision notices, reason codes, preserved originals, independent audit, human review, and repair.
Searches, prompts, voice, gaze, typing rhythm, wearables, and neural signals can be converted into theories about emotion, attention, health, belief, or intent.
Data minimization, purpose limitation, scientific validation, non-retaliatory refusal, notice, and appeal.
Risk scores can convert probability into suspicion, eligibility, discipline, or punishment while hiding uncertainty and feedback loops.
Uncertainty disclosure, fresh evidence, proportional use, independent validation, human authority, and due process.
Generated faces, voices, witnesses, documents, and public participants can borrow human trust and also make authentic evidence easier to deny.
Disclosure, provenance, authenticated official channels, corroboration, and protection for lawful anonymity.
Systems can alter tone, timing, framing, and pressure in response to a person’s behavior and inferred vulnerability.
Meaningful consent, vulnerability limits, transparent purpose, outcome audits, and bans on exploitative optimization.
Fluent, persistent systems can become advisors, companions, interpreters, and identity validators while providers retain control of memory and access.
Role clarity, anti-sycophancy design, visible memory, easy exit, human alternatives, and crisis escalation.
Generated and coordinated identities can simulate social proof, expertise, outrage, conversion, and majority opinion.
Network-level analysis, rate limits, agent disclosure, provenance, crisis circuit breakers, and public verification.
Defending against foreign manipulation can become a pretext for ideology inference, permanent domestic surveillance, and treating lawful dissent as hostile influence.
Behavior-focused investigation, protected dissent, independent oversight, media literacy, authenticated communication, and human control.
Surveillance, arbitrary classification, censorship, coercive uncertainty, and unreviewable force can produce self-censorship, epistemic stress, distrust, and escalation feedback loops.
Protected dissent, de-escalation, open evidence, visible authority, independent review, reintegration, and meaningful appeal.
