1,011 alerts fired in 33 days for fraudulent vessel documentation. Every single one carried confidence 1.000 and robustness score 1.000. Zero have been reviewed.
The Setup
The fraudulent_documentation motif tracks behavioral sequences in the vessel graph where declared cargo, route, or identity information does not cohere with kinematic and structural evidence. It debuted on June 18, 2026 with 116 alerts in its first week. By the week of July 13, it was producing 257 alerts in seven days.
Total as of July 21: 1,011 alerts against 925 unique vessels. The motif fired at maximum confidence on every detection — no probabilistic spread, no sub-threshold events. This is the fourth motif class in the current queue, and the only one whose entire alert population has appeared at 1.000 confidence since the first detection.
The Chain
The weekly cadence shows consistent growth without plateau: 116 alerts (week of June 15), 192 (June 22), 191 (June 29), 195 (July 6), 257 (July 13), with 60 alerts already logged in the first two days of the current week. The velocity has not stalled.
The other motifs in the same queue operate differently. AIS_manipulation_dark carries avg confidence 0.826 across 44,640 alerts — a probabilistic spread across a large vessel population. Ship_to_ship_transfer and AIS_manipulation_spoof fire at 1.000 confidence but represent structurally distinct behavioral patterns. Fraudulent documentation is distinctive: every alert is a maximum-confidence call, and the weekly rate has climbed from 116 to 257 in five weeks.
925 unique vessels are flagged. Entity IDs are formatted as IMO numbers, meaning these are tracked, identifiable vessels, not anonymous MMSI registrations that rotate frequently.
The combined compliance queue now stands at 1,011 (fraudulent_documentation) + 44,640 (AIS_manipulation_dark) + 36,476 (ship_to_ship_transfer) + 5,506 (AIS_manipulation_spoof) + 18 (destination_laundering) = 87,651 unreviewed alerts across five motif classes.
The Implication
A 1.000 confidence score means the detection algorithm produced no borderline cases. Either the underlying behavioral signature is clean enough that every detected instance is unambiguous, or the motif is young enough that it has not yet encountered the edge cases that generate sub-max detections. Both interpretations carry the same operational consequence: 1,011 alerts sitting unreviewed, against 925 identifiable vessels still operating in the maritime system.
Fraudulent documentation is the smallest motif class by count and the newest by age. It is also the fastest-growing. At the current rate of 200+ alerts per week, it will surpass AIS_manipulation_spoof's total (5,506) within approximately 23 weeks if the rate holds.
The directional claim: fraudulent documentation alert velocity is increasing and the queue will widen unless review capacity scales to match it. The 33-day trend shows no sign of natural deceleration.
What to Watch
Whether the current week closes above 200 alerts. The partial week (60 alerts in two days) is on pace for approximately 210, consistent with the prior weeks. If it closes materially below that, the growth rate may be moderating. If it exceeds 300, the trajectory accelerates.
The first acknowledged alert is a meaningful threshold. Zero acknowledgments across all five motif classes means no operator has interacted with any detection. The first review would confirm the queue is monitored at all.
Limitations
The fraudulent_documentation motif is 33 days old. Its behavioral signature has not been validated against confirmed ground truth. Maximum confidence on every alert may indicate a precise, well-calibrated pattern — or it may indicate the motif fires on a narrow rule set that has not yet encountered the edge cases that would generate sub-max outputs. Without reviewed alerts as a reference set, the 1.000 confidence score describes the algorithm's certainty, not a confirmed accuracy rate. Treat the 925-vessel figure as a queue for investigation, not as a confirmed count of fraud events.
Data as of July 21, 2026, 11:20 UTC. Sources: motif_alerts table, fraudulent_documentation behavioral motif classifier.