Configure personal AI interaction preferences

Core meaning

Capability purpose: Make interaction style explicit and revisable without confusing preference with truth, permission, or execution policy.
Meaningful outcome: An active, versioned preference set can guide eligible interactions within its declared scope.
Boundaries — includes: tone, response style, verbosity, formatting, ask-versus-assume posture, explanation depth, evidence presentation, proposal cadence, scope, provenance, and version.
Boundaries — excludes: personal-AI identity, owner profile facts, execution permissions, standing rules, model routing, and request-specific instructions.
Terms and concepts: An interaction preference shapes presentation or collaboration behavior; it does not authorize an action or make a statement true.

Actor: The owner or an authorized interface acting on an explicit owner choice.
Command / intent: Configure how the personal AI should communicate and collaborate in a defined scope.
Current state: An owner/Core context, supported preference vocabulary, and current preference version exist.
Policies / invariants: Changes are explicit and attributable; observation alone does not silently convert into a durable preference; preferences cannot broaden data access, execution authority, or truth status; narrower request-scoped instructions remain distinguishable.
Transition: Validate the selected values and scope, record the decision and provenance, supersede prior values where applicable, and expose the active version.
Result: An active, rejected, or superseded interaction-preference version.
Events / effects: May influence response presentation, clarification behavior, proposal cadence, and evidence display.
Unknowns: Universal precedence among global, role-scoped, mode-scoped, and request-scoped preferences is not established.

Interaction preferences must not override execution policy, standing-rule authority, evidence requirements, or owner correction.
Generated or inferred preference suggestions require an explicit governance path before becoming durable defaults.

The minimum portable preference vocabulary across personal-AI implementations remains undefined.

Statement	Semantic class	Evidence status	Source / note
Personal-AI interaction preferences influence communication and collaboration but remain separate from truth and operational permission.	rule/invariant	sourced	D06.