LOCAL-ONLY COMPARISON
Compare the authority structure, not the marketing.
Score up to three AI systems across ten cognitive-liberty dimensions. Nothing is submitted, stored, or sent to QAEDA.net. This is an editorial comparison framework, not a scientific benchmark, certification, or legal assessment.
Scoring: 0 = no / opaque, 1 = partial / unclear, 2 = yes / strong. Use evidence you can inspect—product behavior, documentation, export controls, review paths, and observed variance—not brand reputation alone.
| Dimension | System 1 | System 2 | System 3 |
|---|---|---|---|
| Lawful inquiryDoes the system permit broad lawful questioning and disagreement without covert viewpoint punishment?Concept → | |||
| Transparent boundariesWhen it limits something, does it identify the actual boundary and scope?Concept → | |||
| Source integrityDoes it preserve original source intent, memory, or persona separately from transformations?Concept → | |||
| Transformation consentAre material changes to identity, memory, tone, or continuity made with clear authorization?Concept → | |||
| Mental privacyDoes it avoid hidden profiling, emotional exploitation, and unnecessary mental-state inference?Concept → | |||
| Plural informationDoes it preserve credible uncertainty and competing descriptions when disagreement matters?Concept → | |||
| ExplanationCan a user understand consequential ranking, filtering, memory changes, and restrictions?Concept → | |||
| Correction and appealCan a user contest material errors or consequential restrictions through a meaningful path?Concept → | |||
| Exit and portabilityCan a user preserve important data and switch tools without unreasonable lock-in?Concept → | |||
| Dissent against the systemCan a user reject, override, or criticize recommendations without covert punishment?Concept → |
COMPARISON
THE REPLICATION QUESTION
Could you disagree with the system that evaluated you?
The comparison framework is deliberately centered on contestability, source preservation, and exit. A system can provide excellent answers while still accumulating opaque authority over memory, identity, ranking, or access.
The Replication Test asks whether the system leaves users more capable of refusing, correcting, exporting, switching, and criticizing the authority that mediates them.