Output-to-Action Traceable

error

output_action_traceable

ai-supervision auditability pure_temporal ai-governancesupervisoryioscorecordkeepingrag-redtier-2-4

Every consequential action is traceable to a logged AI output that preceded it.

Formula

G((Release ∨ Deploy) → P(Output ∧ logged))

Why it matters

IOSCO Table 6.3: 'What processes ensure traceability between AI outputs and final actions taken?' The action must link back to the recorded output that drove it.

Examples

passes Output(logged) → Release

fails a Release with no logged AI output in its history

Use it

ponens policies add output_action_traceable --into ./trace.json
ponens trace check ./trace.json