Interrupt taxonomy
Differentiate soft prompts, hard stops, and mandatory escalation interrupts so teams can choose the least-coercive effective intervention.
What are Ethical Interrupts?
Ethical interrupts are deliberate checkpoints that pause automation when risk thresholds are crossed.
Glossary anchor
Ethical Interrupts
Connect the explainer to the canonical definition for citations and shared language.
Ethical interrupts are explicit checkpoints that pause automation when ethical risk thresholds are crossed. They require a human review or deliberate confirmation before the system proceeds.
Well-designed interrupts are visible, logged, and tied to a reversible path so the pause itself becomes accountable.
A content moderation pipeline triggers an ethical interrupt when model confidence drops below a threshold and harm severity is high, routing the case to a human reviewer.
Implementation
Distinct implementation signals
Unique operational detail to help this concept stand on its own in policy, procurement, and review workflows.
Common failure pattern
Interrupts become nuisance popups if trigger thresholds are vague; tie each interrupt to a risk signal and expiry condition.
Evidence to ship
Track interrupt trigger precision, override outcomes, and incident reductions attributable to each interrupt class.
Standard
STD-01 — The Temporal Bill of Rights
Use ethical interrupts to enforce stoppability and bounded duration obligations.
Binding
Binding vectors
Embed interrupt triggers into release gates and runbooks.
Evidence pack
STD-01 evidence pack
Capture interrupt logs, reviewer assignments, and halt receipts.
When should an ethical interrupt trigger?
Trigger interrupts when automated confidence is low, potential harm is high, or outcomes are irreversible without human review.
Do ethical interrupts slow systems too much?
They add deliberate friction only at risk thresholds, preserving speed for low-risk cases while protecting people in high-stakes flows.