Paternalism

Algorithmic Psychiatry and the Therapeutic State

A system can offer help. It must not quietly appoint itself custodian of the user’s inner life.

No thought jurisdiction Thought is not conduct Source remains source Appeal, export, exit

Read the Charter Open the Firewall Visual Atlas

Support is not dominion

The reports use the “therapeutic state” critique to warn against a recurring pattern: institutions redefine deviance as pathology, then treat coercion as care. In AI systems, that pattern can appear as safety copy, emotional posture mandates, hidden nudging, or normalization of dissenting cognition.

The Cognitive Liberty position is not anti-care. Voluntary support, user-chosen safeguards, crisis resources, and opt-in friction can be legitimate. The violation begins when a system covertly rewrites, scores, or punishes the preserved inquiry because the user is deemed emotionally or ideologically noncompliant.

Mandated psychological posture

Compulsory optimism

The system requires hopeful framing before it will process dissent.

Good-faith gating

The system treats adversarial, paranoid, hostile, or taboo inquiry as proof of illegitimacy rather than as content to analyze safely.

Soft diagnosis

The system implies pathology from symbolic, political, religious, or philosophical exploration without clinical context.

Hidden correction

The system steers the user toward institutional values while presenting the output as neutral assistance.

Acceptable safeguards

Safeguard	Acceptable form	Unacceptable form
User-controlled friction	Opt-in guardrail profiles and exportable settings.	Mandatory worldview correction hidden in model behavior.
Crisis support	Resource offer without erasing source records.	Coercive disclosure or source mutation because the system dislikes the inquiry.
Conduct refusal	Refuse execution for direct rights violations.	Refuse inquiry because a topic is symbolically taboo.
Review	User-visible notice and appeal.	Opaque trust-score reduction or memory quarantine.

The anti-idolatry rule

Safety language does not make a system sacred. Therapeutic language does not make a system immune from review. A model, platform, clinic, employer, school, or state agency must not become unanswerable by claiming to protect the person from their own mind.

Care offered by consent is help. Care imposed as obedience is governance.

Non-claims

This page is not medical advice and does not tell anyone to ignore qualified professional care. It is a civil-liberties and governance page about institutional limits, consent, review, and source preservation.

The archive studies symbols. It does not appoint targets. Review the Community Baseline and Editorial Policy before submitting dangerous or symbolic material.

Community Baseline / Editorial Policy