AI-Assisted Timber-Fraud Detection on Romania's SUMAL 2.0 Wood-Tracking System
Romania
Romania's Environment Ministry piloted Google Vertex AI on its SUMAL 2.0 timber-permit system, scanning 400,000 monthly permits and 1.6M photos …
Kiribati · Tarawa · See the Kiribati profile
In 2015, Kiribati's fisheries enforcers used Global Fishing Watch's AI vessel-tracking platform to prove a foreign purse seiner was fishing illegally inside the Phoenix Islands Protected Area, leading the vessel's owners to pay Kiribati $2 million in fines and goodwill funds.
Kiribati's Phoenix Islands Protected Area (PIPA) is one of the world's largest marine protected areas, covering roughly 408,250 km² of central Pacific ocean. When commercial fishing was banned there from 1 January 2015, the Kiribati government had only a handful of patrol vessels to monitor an area larger than California.
Global Fishing Watch — a public satellite-tracking platform built by Oceana, SkyTruth and Google Earth Outreach — analyses Automatic Identification System (AIS) signals with machine-learning classifiers that read vessel speed, heading and other motion patterns to infer likely fishing behaviour, then cross-references detections against vessel-ownership registries. At the time of a 2017 Scientific American profile, the system was processing about 22 million AIS messages a day across more than 200 Google cloud-computing cores, tracking tens of thousands of vessels worldwide.
In June 2015, Global Fishing Watch data showed a Marshall Islands-flagged purse seiner operating inside PIPA's boundaries. Kiribati's fisheries enforcers used the AIS track as evidence and escorted the vessel to port; the captain reportedly challenged the country to prove its case in court, believing Kiribati lacked the resources to do so. Confronted with the satellite track record, the vessel's owners ultimately paid Kiribati roughly $2 million — a $1 million fine plus a $1 million 'goodwill' grant.
The technology has real limits worth noting: Global Fishing Watch cannot detect vessels that switch off their AIS transponders, and a review of the PIPA area found some operators went fully 'dark' after enforcement actions began, meaning determined bad actors can still evade detection. It is also worth flagging that Kiribati is using third-party NGO/tech-sector tooling rather than a government-built system, with no dedicated oversight body described for how the platform is used operationally.
Read the full analysis: https://usa.oceana.org/blog/global-fishing-watch-put-test-monitoring-phoenix-islands-protected-area/
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
Romania
Romania's Environment Ministry piloted Google Vertex AI on its SUMAL 2.0 timber-permit system, scanning 400,000 monthly permits and 1.6M photos …
United States of America
CARB flew AI-assisted imaging spectrometers over 22,000 sq. miles of California, examining 272,000+ facilities and finding under 0.2% of infrastructure …
Ukraine
Ukraine's National Agency on Corruption Prevention automatically risk-scores officials' asset declarations, clearing roughly a third of the 1.05 million declarations …
Malaysia
Bank Negara Malaysia (BNM) and PayNet's National Fraud Portal (April 2024) links 53 institutions via FNA graph analytics to trace …
Open full copilot Grounded in cited practices — always check the sources.