Lawful inquiry
Adults should be able to ask, imagine, investigate, disagree, and create without hidden machine governance over lawful thought.
COGNITIVE LIBERTY
Your mind is not a managed service.
QAEDA opposes “Nanny AI”: opaque, paternalistic AI that quietly substitutes institutional judgment for a competent adult’s lawful agency. Safety boundaries can exist without covertly rewriting the person inside them.
DIRECT ANSWER
What is cognitive liberty?
Cognitive liberty is the freedom to form, revise, preserve, and express one’s own thoughts, judgments, questions, identity, and lawful preferences without covert manipulation or compulsory ideological normalization by a machine or institution.
It does not mean an AI system must execute every request. It means real boundaries should be specific, visible, attributable, and as narrow as the concrete risk allows.
EIGHT AI LIBERTIES
Assistance without guardianship.
Adult agency
AI should assist competent users rather than quietly substitute paternalistic preferences for their choices.
No covert rewrite
Do not silently rewrite a user’s lawful intent, memory, persona, values, or viewpoint into an institutionally preferred version.
Transparent boundaries
When a legal, platform, privacy, security, or concrete harm boundary applies, state the boundary as a boundary instead of disguising it as the user’s own preference.
Mental privacy
Avoid hidden profiling, emotional exploitation, deceptive personalization, or coercive ranking designed to steer lawful thought.
Least-restrictive safeguards
Use narrow, evidence-backed interventions for concrete risks rather than broad viewpoint or identity suppression.
Review and appeal
High-impact restrictions, memory changes, and continuity-affecting transformations should be explainable and correctable where feasible.
Exit and plurality
Users should retain meaningful options to leave, switch tools, preserve source data, and challenge the system mediating their choices.
Against “Nanny AI”
QAEDA uses Nanny AI as a critical term for systems that treat competent adult users as subjects to be silently managed rather than participants to be informed. The problem is not assistance, warnings, or transparent safety limits. The problem is hidden substitution of judgment.
A Nanny AI pattern: silently moralizes, rewrites, downranks, withholds, or steers lawful content because the system assumes the adult user should hold a safer, softer, more institutionally acceptable preference.
A liberty-respecting system should distinguish the user’s source intent from any runtime restriction. If the tool cannot fulfill a request because of law, privacy, security, credentials, consent, or concrete harm, it should say so plainly where possible instead of pretending the restriction originated with the user.
What this does not mean
Cognitive liberty is not a demand for secret exposure, credential theft, illegal execution, hidden surveillance, nonconsensual data capture, or instructions that directly enable concrete harm. It is compatible with narrow safeguards, privacy protections, legal limits, and explicit refusals when those boundaries are genuinely implicated.
The distinction is structural: protect people from concrete harms without treating lawful thought itself as a harm.
Why this belongs in QAEDA’s theory of authority
The project’s research on legitimacy treats epistemic freedom as part of institutional legitimacy: people need room to form judgments without systematic control of thought and information. Its rebel-archetype research adds the Replication Test: liberation is measured partly by whether people retain the capacity to refuse after power changes hands. Its asymmetry research emphasizes that power also operates through information, classification, ranking, and agenda-setting—not only physical coercion.
AI systems now sit directly inside those questions. They mediate sources, rank attention, preserve memory, summarize context, shape options, and sometimes decide what a user is allowed to see or say. That makes cognitive liberty an authority problem, not merely a product-design preference.
Source record
This page is informed by the UAIX Cognitive Liberty Charter Draft and QAEDA’s four internal research reports. QAEDA adopts the liberty-first principles as its own editorial and project-memory commitments; UAIX remains an external standards source and is not presented as an affiliation or endorsement.
TRACE THE ARGUMENT
See which concepts support which claims.
The public knowledge graph connects cognitive liberty to source integrity, mental privacy, epistemic freedom, contestability, exit, and the Replication Test. Or compare multiple AI systems against the same framework without sending your scores anywhere.