RFC-034: Domains of Concern for Mediated Intelligence — 6. Domains as Constraint-Equivalence Classes

AIGP Specification › RFC-034: Domains of Concern for Mediated Intelligence › 6. Domains as Constraint-Equivalence Classes

← 5. Conceptual Model · Section index · 7. Relationship to Governance →

6. Domains as Constraint-Equivalence Classes

Two use cases belong to the same Domain of Concern when they require materially similar:

evidence requirements,
uncertainty treatment,
authority boundaries,
observer competence,
artifact admissibility rules,
risk thresholds,
concern variables,
calculation patterns,
escalation rules.

For example, the following may all belong to the Mediation Domain of Autonomous Systems:

military drone action selection,
cloud auto-remediation,
automated trading halt,
medication dispensing robot,
self-driving vehicle collision avoidance,
industrial robot shutdown action.

These systems exist in different industries, but they share similar mediated intelligence constraints:

Can the system act?
Was it authorized to act?
Was the action inside its operating envelope?
Was the environment sufficiently known?
Was human control required?
Was the action reversible?
Was the consequence acceptable?
Can the decision be traced?
Can future autonomy be adjusted?

← 5. Conceptual Model · Section index · 7. Relationship to Governance →