Direct answer
A cognitive-liberty system encourages voluntary knowledge sharing, protects inquiry without prejudgment, exposes legitimate minority viewpoints, tests AI for representational blind spots, and makes consequential discoverability decisions contestable. Personal privacy remains available to those who want it.
Eight participation-first principles
Ask without prejudgment
A lawful research question does not imply endorsement or criminal intent.
Share knowledge voluntarily
People should be free to publish ideas, evidence, and experience and decide how their work may be reused.
Discover dissent
Make relevant uncommon and minority viewpoints accessible for examination, without granting equal evidentiary credibility.
Represent humanity in AI
Test model datasets, retrieval and answers for omissions of languages, cultures and real disagreement.
Keep primary sources visible
Use hyperlinks, citations and independent publishers to restore pathways from answers to original work.
Challenge gatekeeping
Provide transparent policies, audit methods and corrections for demonstrable exclusion.
Practice evidence-based disagreement
Explain uncertainty, consensus and dissent accurately; never confuse availability with validation.
Make privacy a choice
Protect confidentiality when desired, without making social withdrawal the purpose of cognitive liberty.
The expanded 20-article charter explains how these principles apply to lawful inquiry, AI training and retrieval, original-source discoverability, and meaningful appeals. The implementation guide sets out tests that publishers and systems can perform. Explore the research collection.
Eighteen continuing safeguards
Freedom of inquiry
Lawful thinking, questioning, imagining, researching, creating, and disagreement should not be subject to hidden machine governance.
Curiosity is not intent
A topic, search, book, or question should not be treated as proof of endorsement, character, or preparation for harm.
Mental privacy
Neural data, search histories, AI conversations, and sensitive cognitive inferences deserve heightened protection.
Search privacy
Private queries should be minimized, separated from identity, protected against dragnet access, and excluded from unrelated eligibility decisions.
Adult agency
Adults should retain meaningful control over lawful choices and the tools, assistants, and information environments they use.
No covert manipulation
Systems should not secretly exploit psychological vulnerabilities or reshape lawful viewpoints through concealed optimization.
Identity integrity
User-selected personas, memories, preferences, and self-descriptions should not be silently rewritten or merged with hidden inference.
Consent and provenance
Material transformations should be explicit, attributable, reversible where feasible, and preserve the distinction between source and derivative.
Transparent boundaries
Legal, platform, runtime, or security constraints should be disclosed as constraints—not disguised as the person’s own preference.
Least-restrictive safety
Target demonstrable harmful conduct or narrowly defined capability rather than broad categories of lawful thought.
Safety without surveillance
Prefer local processing, anonymous credentials, short retention, private retrieval, and aggregate measurement over population-scale logs.
Epistemic due process
High-impact restrictions and classifications should provide notice, reasons, access to relevant data, correction, and meaningful appeal.
Anti-ratchet design
Exceptional power needs purpose limits, automatic deletion, independent authorization, public statistics, technical separation, and hard sunsets.
Public accountability
Systems that shape cognition should make their claims, boundaries, changes, error rates, and enforcement processes inspectable.
Anonymity by default
Lawful access and participation should not require a reusable legal identity when a minimal, unlinkable proof is sufficient.
Network openness
Routing, hosting, protocols, models, payments, and app distribution should remain diverse enough to resist a single censorship switch.
Distributed memory
Archives, evidence, culture, and journalism should be preserved across institutions, languages, formats, and legal jurisdictions.
Worldwide rights floor
Cross-border governance should protect thought, privacy, expression, association, due process, and access without imposing one centralized orthodoxy.