A 5th Motif Class Entered the Queue at Maximum Confidence. The Other 4 Have Produced 66,973 Alerts in 88 Days. Zero Have Been Reviewed.

On June 24, the behavioral graph produced its first destination_laundering alert. By June 29 it had produced 13. Each carries a robustness score of 1.000. None have been reviewed.

That is the full story of the fifth motif class. It entered a queue with 66,973 other unacknowledged alerts and became indistinguishable from them.

The Setup

The compliance queue as of July 3: 66,986 alerts, five distinct motif classes, zero acknowledgments. The oldest unacknowledged alert dates to April 6 — 88.2 days ago. All 66,986 are classified high severity. 11,469 new alerts arrived in the last 7 days.

The queue is dominated by AIS_manipulation_dark (34,358 alerts, 51.3%) and ship_to_ship_transfer (31,302, 46.7%). AIS_manipulation_spoof holds 898 alerts but went silent on May 15 — 49 days ago. fraudulent_documentation adds 415. And now destination_laundering: 13 alerts, all from a 5-day window, then nothing.

The Chain

destination_laundering describes vessels that misrepresent their declared destination — a technique that obscures port calls, bypasses destination-aware screening, or launders routing history before entering regulated terminals. The robustness score of 1.000 means the temporal graph found unambiguous path sequences, not probabilistic inferences. These are complete chains.

The motif fired for 5 days and went quiet. Whether that is a classifier adjustment, a vessel cohort that changed behavior, or the end of a short detection window is not determinable from the alert record alone. What is determinable: 13 maximum-confidence detections of a new evasion technique are currently sitting unread at positions 66,974–66,986 of an 88-day-old queue.

AIS_manipulation_spoof is the prior example. It fired its last alert on May 15 and has 898 unacknowledged entries. Whether those represent active evasion events or a classifier artifact that has since been retired — there is no adjudication record either way.

The Implication

When a queue runs 88 days without a single acknowledgment, the classifier confidence score stops being an operational lever. A detection at robustness 1.000 and a detection at robustness 0.3 sit in the same unreviewed pile. The queue has stratified by motif volume — AIS_manipulation_dark dominates because it fires most — but not by confidence or novelty.

destination_laundering entered the detection system at the highest possible confidence level and immediately became invisible. The queue design does not surface high-confidence new-motif detections above the existing backlog.

What to Watch

Whether destination_laundering resumes firing. A 5-day burst followed by silence is ambiguous — classifier tuning artifact or leading edge of a pattern. No adjudication status means no answer yet.

The AIS_manipulation_spoof thread: 898 alerts, silent since May 15, zero adjudications. Whether those are stale entries or unresolved evasion events is a separate question from the current queue depth.

Queue growth rate: 11,469 alerts in the last 7 days. At that pace the backlog crosses 80,000 before end of July.

Limitations

The 90-day query window captures alerts from April 6 forward; older backlog is not reflected in these counts. Robustness scores are derived from path-edge patterns in the temporal graph — a score of 1.000 indicates path completeness, not that the underlying event was materially evasive. Severity is uniform at high across all five motif classes; this may reflect a classification default rather than differentiated triage.

Data: motif_alerts, queried 2026-07-03. Alert counts bounded to detections since 2026-04-06.