Open Autonomous Intelligence Initiative

One of the most important — and least understood — aspects of trustworthy autonomous intelligence is explicit event modeling.

In OAII, systems do not reason about people they reason about events that are:

• observable
• contextual
• policy‑mediated
• explainable

Enumerating events upfront is not bureaucracy. It is how we prevent silent inference, hidden judgment, and ungovernable behavior.

What follows is a canonical OAII PER event list suitable for edge‑primary aging‑in‑place systems. It is intentionally conservative, extensible, and human‑reviewable.

---

How to Read This List

Each event includes:

• Event ID – stable, machine‑readable identifier
• Description – plain‑language meaning
• Detection Basis – what signals may contribute
• Policy Notes – constraints on use and escalation

No event implies diagnosis, intent, or compliance.

---

Category A: Arrival & Departure Events

PER.A.001 – PrimaryUserArrived

• Description: Primary user has arrived home
• Detection Basis: Door motion + known tracker presence
• Policy Notes: No action by default; establishes temporal context

PER.A.002 – PrimaryUserDeparted

• Description: Primary user has left home
• Detection Basis: Door motion + tracker absence
• Policy Notes: Clears active routines; no alerts

PER.A.003 – ArrivalWithoutKeys

• Description: User arrived without keys detected
• Detection Basis: Arrival + missing key tracker
• Policy Notes: May trigger gentle reminder only if enabled

PER.A.004 – ArrivalWithoutPhone

• Description: User arrived without phone detected
• Detection Basis: Arrival + missing phone tracker
• Policy Notes: Logged; no escalation by default

PER.A.005 – UnexpectedEntryDetected

• Description: Entry detected without known tracker
• Detection Basis: Door motion + unknown presence
• Policy Notes: Policy‑gated; never auto‑accusatory

---

Category B: Object Placement & Routine Completion

PER.B.001 – KeysPlacedInDesignatedLocation

• Description: Keys placed in expected location
• Detection Basis: Key tracker proximity
• Policy Notes: Positive confirmation only

PER.B.002 – PhonePlacedInDesignatedLocation

• Description: Phone placed in expected location
• Detection Basis: Phone tracker proximity
• Policy Notes: Positive confirmation only

PER.B.003 – RoutineCompletedSuccessfully

• Description: Expected arrival routine completed
• Detection Basis: Arrival + object placement
• Policy Notes: No alert; may suppress reminders

PER.B.004 – RoutineIncompleteAfterArrival

• Description: One or more expected steps missing
• Detection Basis: Time window expired
• Policy Notes: Reminder only if explicitly allowed

---

Category C: Activity Presence & Absence

PER.C.001 – MovementDetectedAfterArrival

• Description: Normal movement detected post‑arrival
• Detection Basis: Motion sensor
• Policy Notes: Contextual only

PER.C.002 – ProlongedInactivityDetected

• Description: No movement for extended period
• Detection Basis: Absence of motion
• Policy Notes: Escalation strictly policy‑controlled

PER.C.003 – MovementResumedAfterInactivity

• Description: Activity resumes following inactivity
• Detection Basis: Motion sensor
• Policy Notes: Clears concern state

---

Category D: Routine Drift & Pattern Change

PER.D.001 – ArrivalTimePatternShift

• Description: Arrival times differ from historical norm
• Detection Basis: Temporal comparison
• Policy Notes: Logged only; no inference of cause

PER.D.002 – IncreasedRoutineCompletionTime

• Description: Routines taking longer than usual
• Detection Basis: Time‑to‑completion trend
• Policy Notes: Requires sustained pattern

PER.D.003 – DecreasedRoutineFrequency

• Description: Routine occurring less often
• Detection Basis: Event frequency analysis
• Policy Notes: Never diagnostic

---

Category E: Forgetfulness‑Related Events

PER.E.001 – MissedMedicationEvent

• Description: Expected medication event not detected
• Detection Basis: Time window + absence of signal
• Policy Notes: Requires explicit consent and configuration

PER.E.002 – RepeatedMissedReminder

• Description: Same reminder missed multiple times
• Detection Basis: Event repetition
• Policy Notes: Human review recommended

---

Category F: Boundary & Privacy Events

PER.F.001 – MonitoringPausedByUser

• Description: User explicitly paused monitoring
• Detection Basis: User action
• Policy Notes: Absolute override

PER.F.002 – MonitoringResumedByUser

• Description: User re‑enabled monitoring
• Detection Basis: User action
• Policy Notes: Resets context

PER.F.003 – EventSuppressedDueToPrivacy

• Description: Event intentionally not observed
• Detection Basis: Privacy policy
• Policy Notes: Logged as reminder of non‑observation

---

Category G: Caregiver & Social Context

PER.G.001 – KnownCaregiverArrival

• Description: Authorized caregiver arrived
• Detection Basis: Known tracker
• Policy Notes: Informational only

PER.G.002 – RepeatedUninvitedVisitObserved

• Description: Pattern of unexpected arrivals
• Detection Basis: Temporal repetition
• Policy Notes: Surfaced for human review only

---

Why This Enumeration Is an Ethical Safeguard

Every event above is:

• observable
• explainable
• bounded
• policy‑mediated

None of them:

• assign intent
• score behavior
• diagnose conditions
• infer motivation

This is not a limitation. It is the ethical core of OAII.

---

The Bigger Picture

When autonomous systems reason in terms of explicit events instead of implicit judgments, we gain:

• transparency instead of opacity
• governance instead of trust‑me
• dignity instead of surveillance

If autonomous intelligence is going to exist in our homes, this is the minimum standard it must meet.

---

OAII is not about building smarter machines. It is about building systems we can live with.