Human Factors and Operator Reliability

Quantum platforms are operated by people under uncertainty. Operators interpret ambiguous target health, approve changes, respond to provider incidents, preserve evidence, and decide whether a run can support a claim. Human reliability is part of platform reliability.

flowchart LR
    Interface[Operator interface] --> Decision[Decision]
    Decision --> Action[Action]
    Action --> Platform[Quantum platform]
    Platform --> Feedback[Telemetry and evidence]
    Feedback --> Interface

Error-prone conditions

NASA human-factors guidance emphasizes structured analysis of human contribution to incidents. That is useful in quantum operations because the right question is usually not “who made the mistake?” but “what interface, procedure, context, or organizational condition made the mistake likely?” [R173]

mindmap
  root((Operator reliability))
    Interface clarity
    Alert quality
    Handoff discipline
    Procedure quality
    Training drills
    Fatigue controls
    Blameless review

Interface design

Operator interfaces should be organized around decisions, not raw databases.

flowchart TB
    DecisionNeed[Decision need] --> Context[Relevant context]
    Context --> Status[Current status]
    Status --> Risk[Risk indicators]
    Risk --> Actions[Allowed actions]
    Actions --> Consequences[Expected consequences]

For example, a target page should first answer whether the target is usable for a workload class. Raw calibration tables remain available, but they should not be the only interface.

Alert design

Alerts should map to actions.

flowchart LR
    Signal[Signal] --> Rule[Alert rule]
    Rule --> Actionable{Actionable?}
    Actionable -- yes --> Page[Page]
    Actionable -- no --> Dashboard[Dashboard or report]

Handoffs

sequenceDiagram
    participant Outgoing
    participant Incoming
    participant Runbook
    participant Evidence
    Outgoing->>Incoming: current state and risks
    Outgoing->>Runbook: active procedure step
    Outgoing->>Evidence: affected workload IDs
    Incoming-->>Outgoing: repeat back next actions

A handoff should transfer active incidents, affected targets, workload freezes, changes in progress, evidence gaps, vendor tickets, and next decision deadline.

Reviews

NIST log-management guidance reinforces the need for useful, protected, and reviewable logs. Quantum operations need operator-action records, policy decisions, runtime changes, evidence events, and context snapshots that support after-action review. [R172]

flowchart TB
    Event[Event] --> OperatorContext[Operator context]
    OperatorContext --> DecisionPath[Decision path]
    DecisionPath --> Factors[Contributing factors]
    Factors --> Improvements[System improvements]

Training

flowchart LR
    Scenario[Scenario library] --> Drill[Drill]
    Drill --> Observation[Observation]
    Observation --> Feedback[Feedback]
    Feedback --> ProcedureUpdate[Procedure update]
    ProcedureUpdate --> Scenario

Train on stale calibration, provider metadata loss, cost runaway, ambiguous benchmark regression, safety alarm, and evidence preservation failure.

Operating rule

Design quantum operations for real humans. Interfaces, runbooks, alerts, handoffs, and drills are production infrastructure.