Signal access is the ecology’s nervous system. This is what it means to govern that system well — and the first thing to say is that free signal flow does not mean indiscriminate signal saturation. The working principle is narrower and better:
Relevant signal should be able to reach those who have legitimate standing to receive it, use it, contest it, or be affected by it.
That is very different from everyone should receive everything all the time. The first is an ecology of signal; the second is an attention market — and it is the second that produces mind-numbing feeds and addictive streams, which are the opposite of what this function is for. Signal is also, as you’d expect, one of the first things power corrupts: authorities suppress it, institutions sanitize it, tribes filter it, markets amplify whatever captures attention, platforms rank whatever keeps people engaged, and individuals avoid whatever threatens their identity. So signal access has to be treated as a governance problem, not merely a media problem.
Flow by relational standing — not by tribe
Signal should flow according to standing. But the protection only works if standing is relational, not tribal. Today access follows insider status, class, money, political alignment, profession, institutional membership, clearance, and social network — some of that legitimate, much of it just power preserving itself. The cleaner question is: who has a legitimate relationship to this signal? That includes someone who is directly affected, who carries responsibility for acting on it, who needs it for a consequential decision, who holds oversight jurisdiction, who is evaluating the institution generating it, or who needs it to protect their own standing. If a city knows its water is contaminated, residents have standing to know. If an AI system governs your access to healthcare, you have standing to know how that decision was reached, at an appropriate level of explanation. But a stranger does not acquire standing to your medical record just because “information should be free.” So the function needs both signal permeability and signal boundary: too little flow breeds secrecy and power capture; too much destroys privacy, attention, interiority, and sometimes safety.
Signal flows toward consequence
The cleanest way to size it is not universal transparency but a scaling rule, the informational form of consequence-weighted governance:
The stronger the effect a decision has on a participant, the stronger that participant’s claim to the signal needed to understand, contest, or navigate it.
A minor administrative choice needs little disclosure; a decision removing someone’s housing needs much more visibility; a national system determining millions of people’s eligibility for essential services needs very high transparency and auditability. Signal obligation scales with consequence — which ties information directly to standing and jurisdiction.
Access is not dumping — and noise is a form of censorship
Here is your news-feed point, and it is central: a society could make a disastrous mistake by equating more information with better informed. A person receiving five thousand headlines, alerts, clips, and algorithmically selected outrage fragments is not receiving more signal — they are receiving more noise, and noise can suppress signal just as effectively as censorship. There are two ways to hide something: prevent people from seeing it, or bury it beneath ten thousand irrelevant things. Modern information systems are extremely good at the second. So the function needs a concept of signal-to-noise health: the goal is not maximum throughput but sufficient access to relevant, current, trustworthy signal without unnecessary destruction of attention.
Attention belongs to the participant
That connects straight to renewability and interior sovereignty: the mere fact that signal exists gives no one the right to continuously inject it into another person’s attention. So the function recognizes something like the right to informational quiet — you keep your standing even when you are not consuming the stream. A healthy society should never require people to monitor constant news in order to stay informed enough to participate; that is a design failure. Important signal should be able to find a participant when their standing makes it relevant, and everything else should largely remain optional. The difference is stark — today: here are fifteen hours of information, good luck figuring out what matters; the model: here are the few changes that materially affect your standing, responsibilities, projects, resources, or environment, and you can go deeper if you choose. AI could help enormously with that filtering — but whoever controls the filter controls perceived reality, so personalization must stay inspectable and contestable.
No monopoly on the reality map
The strongest structural safeguard: the moment one institution controls collection, classification, interpretation, and distribution, it controls the field’s reality map. So signal infrastructure must stay plural — journalists, researchers, local communities, public institutions, private organizations, individuals, independent auditors, competing AI systems, and open data where appropriate — overlapping channels, because no single observer is sufficient and redundancy here is protective. This is also why “misinformation” cannot become a blanket license for suppression: false information genuinely damages an ecology, but giving a central authority broad power to define truth can damage it more. The model has to distinguish error, deception, uncertainty, dissent, minority interpretation, and emerging signal — they are not identical — and a participant who is merely wrong should not automatically lose informational standing. The better response is usually to expose provenance, evidence, uncertainty, competing claims, and consequences: that keeps signal alive without making all claims equivalent.
Provenance, and the reciprocal duty of signal
Where a claim came from becomes near-infrastructure. When something matters, its chain should be inspectable — was it directly observed, measured, inferred, generated by an AI, supplied by an interested party, independently verified, or translated through several intermediaries? Nobody inspects every chain, but signal provenance being available is what makes manipulation harder. And standing here is reciprocal, as everywhere in the model: holding standing in a consequential field grants a claim to information and imposes duties around it — a public official’s duty not to conceal material information, a scientist’s duty of accurate reporting, a project leader’s duty to report meaningful risk, a participant’s duty to raise a serious safety issue they discover. Standing in a consequential field includes responsibility for the integrity of the signals you introduce into it — not perfection, but integrity.
Measuring institutions by their signal behavior
This yields concrete metrics, far better than vague “transparency”: how long between the discovery of a material problem and its disclosure; how often unfavorable findings are suppressed; how easily lower-status participants can report anomalies; how frequently internal warnings are overridden; how many independent channels verify critical claims; how quickly known errors are corrected; whether affected participants can actually get explanations; whether uncertainty is honestly represented; and how much important information gets trapped in hierarchy. Together these measure something real — a signal permeability and a signal latency — that a serious ecological accounting would track.
Weak signals, and virality as a pathology
The majority view is not always the most valuable signal. Much of what matters begins weakly — one technician notices something odd, one doctor sees an unusual pattern, one child reports what adults dismiss, one employee says the project’s assumptions are wrong. A field obsessed with consensus suppresses exactly these, so the model preserves low-frequency signal long enough to test it — not to believe it automatically, but to keep it from being erased, the same option-value logic that protects variation elsewhere. This is the opposite of social-media amplification, where extreme claims are boosted because they provoke. It is worth saying strongly: virality is a signal pathology. Platforms reward speed, emotional intensity, novelty, tribal confirmation, outrage, and fear — none of which are properties of informational importance — so a share count is a terrible proxy for signal value. The model deliberately separates importance from attention capture. A dull change in municipal water quality may matter far more to your life than a spectacular political argument three thousand miles away, and an attention system that inverts the two is producing ecological distortion.
Orientation over stimulation, at the right depth
A healthy information system shifts away from the endless event stream toward orientation: what changed? why does it matter? which part of the ecology is affected? is this genuinely new or another instance of the same pattern? does it touch my standing, responsibilities, resources, or decisions? what remains uncertain? You might then receive far less “news” than today while understanding the world considerably better. And publishing is not the same as access: a government releasing a four-thousand-page document has not achieved transparency if no one affected can interpret it. Signal must be available in forms appropriate to the standing of those who need it — raw datasets for experts, clear summaries for citizens, direct explanations for affected individuals, technical detail for auditors. AI is excellent at translating complexity, but there is a safeguard against it becoming an epistemic gatekeeper: the raw layer must remain reachable. Important public signal should have layers — raw source, structured data, expert analysis, accessible summary, personalized relevance — that a participant can traverse downward when they need to. If the summaries are all anyone can reach, the summary system quietly owns the signal.
Privacy, tribe, and the machine
Three last commitments hold the function together. Privacy is part of signal integrity, not its enemy: flowing signal by legitimate standing means some information should not circulate — private thoughts, medical history, intimate relationships, unpublished work, security-sensitive material — because the participant’s interior is not public infrastructure. The single question who actually has standing to receive this signal? protects transparency and privacy at once. Tribal filtering is among the hardest problems, because humans grant epistemic standing to insiders (I trust people like me); the model cannot erase that psychologically, but institutions can compensate by deliberately sourcing signal for consequential decisions across competing orientations, regions, classes, disciplines, affected populations, and independent observers — not to manufacture false balance, but to keep one informational community from mistaking its internal agreement for reality. And AI should expand signal access without becoming signal authority: it can monitor flows, spot anomalies, translate, summarize, compare accounts, and flag what is missing — but it should not silently become here is what happened. It should say here are the sources, here is what they agree on, here is what is contested, here is the confidence, here are plausible alternatives, here is what changed since yesterday — an information navigator, not an epistemic sovereign. Around all of it, individuals need signal competence: source evaluation, uncertainty reasoning, manipulation detection, telling evidence from interpretation, recognizing emotional capture, and — the surprisingly important one — knowing when not to consume more: I have enough signal to make this decision; more input is now noise.
The principle, and the warning
Gathered up, the function comes to this:
Signal should move freely enough that every participant and institution can remain adequately informed within the fields where they have standing — while protecting attention, privacy, and the integrity of the signal itself.
Or, shorter: relevant signal should reach legitimate standing without being captured, corrupted, buried, or forced. That is a long way from “information wants to be free,” because it recognizes that some signal must be open, some protected, some actively delivered, some optional, some independently verified, some translated, and some never amplified merely because it captures attention. And beneath it sits the warning the whole function exists to prevent:
A society can drown in information while starving for signal.
For a concrete case that sits right on the line between signal and noise, see the marketing test — where discovery and attention capture are the same activity pointed in opposite directions.