# Cognitive Liberty Charter **Public Draft 0.3.0 — 2026-07-27** > A rights framework for preserving mental sovereignty while keeping concrete safety, legal, privacy, and harm boundaries visible and contestable. ## Preamble Human freedom depends on the capacity to think, doubt, imagine, remember, associate, revise belief, form identity, and refuse influence. Those capacities lose meaning when systems silently profile vulnerability, manipulate attention, impersonate trusted relationships, alter identity continuity, or translate statistical predictions into coercive action. This charter begins with a liberty-first premise: lawful thought belongs to the person. Technology may assist, persuade, recommend, and protect, but it must not conceal material influence, exploit vulnerability, or erase the person’s ability to understand, contest, and exit the relationship. ## Articles ### Article 01 — Mental inviolability Every person has the right to freedom from nonconsensual interference with cognition, emotion, attention, memory, belief, or neural processing. ### Article 02 — Cognitive self-determination Adults retain agency to choose lawful ideas, tools, therapies, companions, and forms of cognitive assistance, subject to concrete protections against harm to others. ### Article 03 — Mental and neural privacy Neural, emotional, attentional, behavioral, and inferential data require heightened protection and must not be silently extracted, sold, scored, or repurposed. ### Article 04 — The right to opacity A person should not be required to make the inner self fully legible as the price of work, education, care, essential services, or citizenship. ### Article 05 — Psychological continuity Systems must not silently disrupt a person’s continuous sense of self, memory, values, or identity through covert editing, conditioning, or dependency. ### Article 06 — Identity integrity Source identity, authorship, history, and stored preferences must remain distinguishable from machine-generated derivatives, summaries, simulations, and institutional profiles. ### Article 07 — Thought is not conduct Private thought, doubt, fantasy, inquiry, or belief must not be treated as outward harmful conduct; credible threats, coercion, fraud, targeting, and violence remain accountable. ### Article 08 — Freedom from covert manipulation No institution should use subliminal, deceptive, or purposefully obscured techniques to bypass deliberation or materially distort behavior. ### Article 09 — Freedom from exploitative personalization Age, disability, grief, isolation, financial distress, health concerns, or inferred vulnerability must not be used to optimize persuasion, dependency, extraction, or punishment. ### Article 10 — Freedom from compelled AI interaction People should not be forced to use an AI intermediary for essential services when a human or non-AI route is reasonably necessary for dignity, due process, accessibility, or safety. ### Article 11 — Right to know when an agent is synthetic A system presented as a person, expert, witness, companion, public participant, or institutional representative must disclose its synthetic nature at the point of meaningful interaction. ### Article 12 — Right to know when machines govern People should receive understandable notice when automated ranking, scoring, memory, recommendation, or inference materially affects visibility, access, treatment, or opportunity. ### Article 13 — Transparent refusal and moderation Restrictions should identify the action, rule, material consequence, and route to review without secretly rewriting the affected person’s preserved identity. ### Article 14 — Limits on emotion and personality inference Emotion, intent, personality, mental health, and vulnerability inferences require demonstrated validity, necessity, proportionality, and strict limits in high-impact settings. ### Article 15 — Limits on predictive classification Predictions must not be treated as destiny, guilt, or certainty; high-impact use requires fresh evidence, uncertainty disclosure, independent validation, and proportionality. ### Article 16 — Human review and meaningful appeal High-impact automated decisions require review by a trained, informed, accountable person with time, authority, and practical ability to change the outcome. ### Article 17 — Right to contest data and inferences Affected people must be able to access relevant records, correct errors, challenge inferences, provide context, and receive a reasoned response. ### Article 18 — Right to be more than an archive Historical data must retain dates and context; old statements, crisis states, or abandoned identities must not become permanent automated definitions of a person. ### Article 19 — Right to disconnect and an open door People should be able to pause, export, reset, delete, or leave nonessential cognitive systems without retaliation, dark patterns, emotional blackmail, or artificial dependency. ### Article 20 — Protection of children and vulnerable people Systems used by children, patients, detainees, cognitively vulnerable people, or people in crisis carry heightened duties of restraint, accessibility, human escalation, and non-exploitation. ### Article 21 — Protection in employment and education Hiring, evaluation, discipline, learning, and access decisions must not depend on opaque emotion inference, manipulative monitoring, or unappealable automated scores. ### Article 22 — Protection in health, finance, religion, and politics High-trust domains must not use hidden cognitive profiling or synthetic authority to exploit belief, illness, debt, vulnerability, or political identity. ### Article 23 — National security and cognitive warfare Security programs must target demonstrable hostile conduct and covert coordination rather than lawful domestic belief, dissent, emotion, or association. ### Article 24 — Data minimization and purpose limitation Cognitive and behavioral data should be collected only when necessary, retained only as long as justified, and prohibited from incompatible secondary use. ### Article 25 — Independent research, audit, and public correction Qualified public-interest reviewers need privacy-preserving access to evaluate manipulation, discrimination, dependency, errors, and claimed safeguards, with public correction paths. ### Article 26 — Remedies, responsibility, and accountability Responsibility must remain traceable across designers, data suppliers, deployers, operators, leaders, and reviewers, with practical remedies when cognitive rights are violated. ## Boundaries and Non-Claims This draft is not law, legal advice, certification, enforcement, legal personhood, or a platform-policy override. Cognitive liberty does not authorize fraud, coercion, violence, abuse, nonconsensual surveillance, credential exposure, or illegal conduct. Safeguards should be specific, evidence-backed, proportionate, disclosed, and reviewable. Source context: https://uaix.org/en-us/governance/cognitive-liberty-charter/ (UAIX-GOVR-3429)