Open Autonomous Intelligence Initiative

How OAII makes context explicit, interoperable, and enforceable

One of the most important — and most easily misunderstood — ideas in the OAII Base Model is the Contextual Axis.

Most AI systems use context. Very few systems model it explicitly.

OAII does, because without explicit structure, context becomes:

• implicit,
• unreviewable,
• non-interoperable,
• and ethically fragile.

This post explains:

1. What a Contextual Axis is conceptually,
2. Why OAII treats axes as first-class objects,
3. How axes can be represented formally (ASN.1), and
4. How the five MVP AxisTypes are used in aging-in-place systems.

1. What Is a Contextual Axis?

A Contextual Axis is a declared dimension along which meaning is interpreted within a World.

An axis answers the question:

“Along what dimension does this observation or event acquire meaning?”

Examples:

• Time of day
• Location within a home
• Observable activity level
• Device health state
• Policy mode (quiet hours, escalation mode)

Crucially, an axis is not the value itself.

It is the dimension, domain, and rules that make values interpretable.

2. Why Axes Must Be First-Class

In many systems, context is:

• buried in feature vectors,
• hard-coded in rules,
• or learned implicitly in model weights.

This creates three problems:

1. No auditability — you cannot ask which context mattered
2. No interoperability — other systems cannot align meanings
3. No ethical boundary — context leaks silently across domains

By contrast, OAII requires axes to be:

• explicitly declared,
• typed,
• bounded by value domains,
• privacy-classified,
• and World-scoped.

This makes context visible and enforceable.

3. Structural Representation (ASN.1)

OAII favors formal, schema-driven representations that are language-neutral and interoperable. ASN.1 is a natural fit because it is:

• explicit about structure,
• widely used in safety- and telecom-grade systems,
• suitable for both wire formats and internal models.

Below is an illustrative ASN.1-style definition of a Contextual Axis.

ContextualAxis ::= SEQUENCE {
    axisId            AxisId,
    axisType          AxisType,
    label             UTF8String,
    valueDomain       ValueDomain,
    unit              UTF8String OPTIONAL,
    frameRef          UTF8String OPTIONAL,
    resolution        REAL OPTIONAL,
    uncertaintyModel  UncertaintyModel OPTIONAL,
    privacyClass      PrivacyClass OPTIONAL
}

AxisType ::= ENUMERATED {
    temporal(0),
    spatial(1),
    activity(2),
    deviceState(3),
    policyContext(4)
}

This structure makes several things explicit:

• what kind of context is involved (axisType),
• what values are allowed (valueDomain),
• how precise it is (resolution),
• how uncertain it may be,
• and how it must be protected (privacyClass).

Nothing is left implicit.

4. The Five MVP Axis Types

The Open SGI MVP constrains the system to five required AxisTypes. Each exists for a distinct reason.

4.1 TEMPORAL Axis

What it represents:

• Time of day
• Duration
• Routine windows

Why it matters:
The same observation has different meaning at different times.

Example:

• Motion in a kitchen at 8am → routine
• Motion in a kitchen at 3am → potentially meaningful

Typical domains:

• Continuous time ranges
• Ordered sets (morning / afternoon / night)

4.2 SPATIAL Axis

What it represents:

• Rooms
• Zones
• Entry/exit boundaries

Why it matters:
Location provides semantic grounding without requiring identity or surveillance.

Example:

• No motion in any room vs. no motion in a specific room

Typical domains:

• Discrete sets of named zones
• Ordered adjacency relationships

4.3 ACTIVITY Axis

What it represents:

• Observable activity categories
• Movement levels
• Interaction presence

Why it matters:
Activity axes allow interpretation without diagnosis or intent inference.

Example:

• “low activity” vs. “high activity”
• “transition” vs. “stationary”

This axis is deliberately non-clinical.

4.4 DEVICE_STATE Axis

What it represents:

• Sensor health
• Device availability
• Data quality

Why it matters:
A system must know when not to trust itself.

Example:

• Prolonged inactivity during sensor outage should not trigger escalation

This axis is essential for graceful degradation.

4.5 POLICY_CONTEXT Axis

What it represents:

• Quiet hours
• Escalation mode
• Maintenance mode

Why it matters:
Meaning is not just observational — it is also governed.

Example:

• The same Event may trigger:
  • no action,
  • local prompt,
  • or caregiver notification
    depending on policy context.

5. Axes as Ethical Infrastructure

By making context explicit and typed, OAII ensures that:

• meaning does not leak across domains,
• policies are enforceable rather than advisory,
• explanations can reference which axes mattered,
• interoperability becomes mapping, not guessing.

Axes are not metadata.

They are ethical infrastructure.

6. Why This Matters for Open SGI and Aging-in-Place

In aging-in-place systems, harm often comes from:

• overgeneralization,
• context collapse,
• silent assumptions.

Contextual Axes prevent this by forcing systems to say:

This interpretation is valid along these dimensions — and no others.

That is the difference between observation and surveillance.

This post explains the role of Contextual Axes in the OAII Base Model and how they are constrained in the Open SGI MVP. Future posts will show how axes participate in Event recognition and Policy enforcement.