Wire Image Intake Gate for News Desks

A pre-publication interrogation gate for photo editors and visual verification leads at small-to-mid-sized newsrooms that receive freelance, partner, or wire images during breaking events. The editor manually enters a small set of external claims attached to an image batch — capture source, time/location confidence, credential presence, chain-of-custody assertions, and any contextual claims from the sender. The product runs AE’s adversarial multi-model debate against a structured constraint set to surface contradiction patterns before publication, then later grades whether the gate’s warnings matched external reality using correction logs, takedowns, provenance updates, and newsroom publication records. The pain removed is not generic deepfake detection; it is the repeated, high-pressure decision of whether an externally supplied image package is safe enough to publish when C2PA signing is becoming common but verification remains unreliable and metadata is often stripped in distribution. AE is suited because this niche needs pre-send challenge logic, explicit failure-pattern taxonomy, and fast reality-grading on concrete publication outcomes rather than subjective ‘trust scores.’

A pre-publication image intake gate is a strong AE wedge because it solves a specific, recurring newsroom verification decision with explainable challenge logic and objective post-publication reality grading.

Killed: evidence_search_exhausted.