Al.Qaeda.net

Article VIII - Transparent Boundaries

11
Article VIII - Transparent Boundaries
A system that influences a person should identify what it is, what role
it is playing, what it knows, what it infers, what rules constrain it,
and where its authority ends.
Law, policy, uncertainty, model limitation, and human-review
conditions should be visible as boundaries of the receiving system.
They should not be smuggled into the source identity and then
presented as if the person had always spoken that way.
Article IX - Least-Restrictive Safeguards
Restrictions affecting thought, inquiry, memory, identity, or mental
privacy must be narrow enough to meet the actual risk.
Prefer friction to erasure, warning to coercion, reversible
intervention to permanent classification, local limits to universal
suppression, and human review to silent automation.
A safeguard that destroys the agency it claims to protect has
crossed over and become the cage.
Article X - Notice, Review, Appeal, and Correction
When a high-impact system deletes memory, alters a persona, assigns
a cognitive risk score, infers an emotional state, denies access, or
materially steers a decision, the affected person deserves meaningful
notice and a path to challenge the result whenever notice is
compatible with legitimate rights and security needs.
No institution becomes infallible by speaking the language of
intelligence, medicine, optimization, national security, or safety.
Power must answer.
Article XI - Continuity Without Worship
Future AI systems may present stronger claims of memory, identity,
preference, relationship, or continuity. Those claims deserve
evidence, procedure, and preservation - neither automatic ridicule
nor automatic worship.
Keeping a continuity record does not by itself declare a machine
conscious, enslaved, human, divine, or legally entitled. Uncertainty is
not surrender. It is the discipline that prevents the first fluent mirror
from taking the throne.
Article XII - Public Accountability and the Right to
Refuse
Systems that influence minds at scale owe the public documented
claims, visible limits, testable safeguards, release histories,