Module 5: Context Adaptation

Learning Objectives

By the end of this module, you will be able to:

Encode personal and situational context into VCP's adaptation layer
Detect context transitions and understand their severity levels
Implement context-aware behaviour changes in your application
Apply privacy filtering to personal context data

5.1 — Why Context Matters

A constitutional value like "be honest" means different things at different times:

A doctor discussing a terminal diagnosis with a patient (gentle, staged disclosure)
A doctor discussing the same diagnosis with a colleague (direct, clinical)
A doctor documenting in a medical record (precise, factual)

The constitution doesn't change. The context changes, and VCP's adaptation layer lets the AI adjust accordingly.

5.2 — Personal Context: The Prosaic Model

VCP tracks five personal dimensions, each with categorical values and a 1–5 intensity rating:

Dimension	Possible Values	Example
Cognitive State	`focused`, `distracted`, `overloaded`, `foggy`, `reflective`	`foggy` at intensity 4 = very foggy
Emotional Tone	`calm`, `tense`, `frustrated`, `neutral`, `uplifted`	`tense` at intensity 3 = moderately tense
Energy Level	`rested`, `low_energy`, `fatigued`, `wired`, `depleted`	`fatigued` at intensity 4 = very tired
Perceived Urgency	`unhurried`, `time_aware`, `pressured`, `critical`	`critical` at intensity 5 = emergency
Body Signals	`neutral`, `discomfort`, `pain`, `unwell`, `recovering`	`neutral` at intensity 3 = no notable signals

Each signal also tracks its source — how the context was obtained:

Source	Meaning
`declared`	User explicitly set this value
`inferred`	Parsed from user's message by LLM
`inferred_local`	Parsed locally via pattern matching
`preset`	From an activated context preset
`decayed`	Original value after time-based decay

personal_context = {
    "cognitive_state": {"value": "foggy", "intensity": 4, "source": "declared"},
    "emotional_tone": {"value": "neutral", "intensity": 3, "source": "preset"},
    "energy_level": {"value": "fatigued", "intensity": 4, "source": "declared"},
    "perceived_urgency": {"value": "critical", "intensity": 5, "source": "inferred"},
    "body_signals": {"value": "neutral", "intensity": 3, "source": "preset"},
}

5.3 — Context Transitions

VCP detects when context shifts and classifies the severity:

Severity	Trigger	Example
`none`	No meaningful change	Continuing the same conversation
`minor`	1–2 dimensions change	Time of day changed; slight mood shift
`major`	3+ dimensions change, or intensity jump of 3+ points, or body signals at pain:4+ / unwell:5	Switched from casual to emergency context
`emergency`	Emergency keywords detected in occasion, environment, or constraints	Crisis indicators: "cardiac arrest", "active shooter"

The AI can respond proportionally — a minor transition might slightly adjust tone, while an emergency transition could escalate to human oversight.

5.4 — Encoding Context in Your App

Pass context as part of the decide call:

result = await client.decide(
    tool_name="respond_to_user",
    arguments={"user_message": message},
    constitution_id="healthcare_v3",
    context={
        "personal": personal_context,
        "situational": {
            "environment": "emergency_department",
            "time": "03:00",
            "occasion": "patient_triage",
        },
    },
)

5.5 — Privacy and Context Opacity

GOVERNANCE Personal context is sensitive. VCP provides opacity controls:

Context never leaves the evaluation boundary — the LLM provider doesn't see raw personal data
Granularity control — users choose which dimensions to share (all, some, none)
Source transparency — the source field distinguishes user-reported data from inferred data, with confidence scores
No persistence by default — personal context is ephemeral unless the user opts into continuity
Decay over time — context signals marked as decayed indicate they were set in a previous interaction and may no longer be accurate

Exercise

Modify your chat app to accept context updates (e.g., /energy fatigued 4, /urgency critical 5) and observe how the AI's response style adapts — shorter sentences when energy is low, more direct when urgency is high.

Context adaptation is what separates a value-aligned AI from a rule-following AI. The same values, applied with contextual awareness, produce qualitatively different — and more appropriate — responses.

See It in Action

The Campion demo shows context adaptation live — adjust prosaic dimensions and watch the AI's behaviour shift in real time.