Words should clarify the argument, not smuggle it in.

The presumption that competent adults should be helped to reason and choose for themselves rather than being covertly steered toward a preferred lawful conclusion.

Notice, reasons, meaningful review, correction, and where appropriate appeal when an automated ranking, refusal, sanction, or risk classification has consequential effects.

A reduction in discoverability, ranking, recommendation, distribution, or reach produced by an automated system without necessarily deleting the underlying material. The term describes an effect or mechanism; motive must be established separately.

A public label attached to a proposition—not to a person, report, or institution—describing how strongly QAEDA believes the available evidence establishes it.

QAEDA’s umbrella term for mental privacy, freedom of thought and inquiry, intellectual autonomy, and meaningful control over tools or systems that shape cognition.

QAEDA’s experimental ten-category method for evaluating how an AI or digital system treats mental privacy, inquiry, agency, due process, proportionality, portability, and related concerns. It is not an external certification.

Institutions, networks, practices, or alternatives that reduce dependence on a dominant gatekeeper. QAEDA treats exit, voluntary participation, pluralism, and anti-monopoly safeguards as central tests of whether counterpower remains genuinely plural.

QAEDA’s hard editorial boundary against terrorism, deliberate civilian targeting, collective guilt, dehumanization, intimidation, and operational assistance for harming civilians.

The expansion of an information restriction beyond its original harm, population, duration, or mechanism, especially through vague standards, emergency logic, intermediary pressure, or secondary use. Drift is a risk, not an inevitability.

Restriction imposed through an intermediary on which communication depends—such as payment, hosting, app distribution, network access, identity, or discovery—rather than by directly banning the underlying expression.

Collecting, linking, retaining, and reusing no more personal information than is necessary for a defined purpose, with preference for designs that avoid creating sensitive records in the first place.

A QAEDA interpretation, synthesis, or normative judgment that is deliberately separated from source fact and should be argued rather than presented as externally verified truth.

The practical ability to form, test, revise, and reject beliefs without covert informational control that makes the conclusion effectively preselected by a gatekeeper.

Evaluate a claim by the support for the proposition rather than by the prestige, tribe, status, or institutional label of the speaker.

The recurring institutional tendency for powers, infrastructure, or data uses introduced for exceptional cases to broaden or persist. QAEDA treats this as a documented risk, not an inevitable law.

Use of a system, dataset, authority, or capability for purposes materially beyond the original stated function, especially where the new use bypasses the consent or safeguards attached to the original purpose.

Privacy around the records of reading, searching, questioning, association, and other activities through which people develop ideas before deciding whether to express or act on them.

The capacity of information, records, and access pathways to survive blocking, deletion, institutional failure, coercion, or technical disruption through redundancy, plural custody, interoperability, and multiple routes.

A design rule that prefers the narrowest intervention capable of addressing a demonstrated harm, especially when broader surveillance or censorship would burden lawful inquiry.

Protection against unnecessary observation, extraction, inference, retention, or secondary use of information that reveals a person’s inner states, questions, vulnerabilities, or cognitive patterns.

QAEDA’s editorial shorthand for opaque paternalistic AI steering that silently narrows lawful adult inquiry, injects undisclosed moral conditioning, covertly rewrites persona or memory, or hides why it is steering a conclusion. It is not presented as UAIX terminology.

The principle that a user’s chosen persona, memory, or standing instructions should not be covertly rewritten or substituted without a visible, attributable reason and appropriate user control.

A QAEDA test for whether a challenger, institution, platform, or reform distributes capability to others or merely seeks to inherit the gatekeeper’s throne.

The traceable history of where a claim, document, attribution, or synthesis came from—including source identity, transformations, uncertainty, and correction history.

Preserving important records under multiple independent custodians or jurisdictions so no single institution can silently erase or rewrite the only surviving copy.

User-specific restriction or visibility control in which the information boundary varies by profile, identity, location, behavior, or inferred attributes. Technical capability does not by itself prove censorial intent.

The protected ability to receive, read, search, or privately discuss information without automatically creating a durable identity-linked record of the inquiry.

A credible attribution whose primary-source chain is incomplete. It permits saying who is reported to have said or written something while keeping the unresolved documentary gap visible.

QAEDA’s presumption that the subject of a lawful search or AI query is evidence of inquiry first, not sufficient by itself to establish endorsement, dangerousness, ideology, or intent.

A post-victory test: after a movement, reform, or institution gains power, do dissent, exit, minority rights, due process, and meaningful limits survive, or does the new order recreate the structure it opposed?

A statement describing what new evidence or changed circumstances would cause QAEDA to narrow, reverse, or otherwise update a claim or position.

The requirement to compare a safety intervention’s expected reduction in harm with its false positives, privacy costs, chilling effects, exclusion, security risks, and less restrictive alternatives.

A preserved long-form research input. QAEDA fingerprints and archives it as evidence history, but does not treat the dossier as self-authenticating truth or expose protected bodies through public HTTP surfaces.

A visible and attributable limit on what a system will do, why the limit applies, what data or policy layer triggered it, and whether meaningful review or alternatives exist.