Ted Hisokawa
March 4th, 2025 08:58
The new AI model developed by AI2, Atlantes will strengthen global fisheries and wildlife surveillance and support conservationists who protect natural resources from illegal activities.
Innovative AI models target illegal fishing
In a groundbreaking development for conservation efforts, researchers have introduced an open source AI model called Atlantes, designed to monitor sailors’ ships around the world. According to Nvidia, the model, developed by AI (AI2) in Seattle, is intended to combat illegal fishing by analyzing 5 billion GPS signals daily.
High accuracy and real-time alerts
Atlantes boasts impressive prediction accuracy of approximately 80% when deciding on vessel activity. It is integrated into AI2’s Skylight Maritime Surveillance Platform, allowing authorities to alert within 15 minutes of detecting potential illegal fishing activities. This ability was demonstrated when the Argentine navy intercepted containers and fined them for illegal fishing following warnings from Skylight.
Atlantes’ Technical Backbone
The AI ​​model, which consists of 4.7 million parameters, is built on an Nvidia H100 tensor core GPU and Pytorch. From January 2022 to June 2024, it processes automated identification system (AIS) data, which is essential for most vessels. The training involved maritime experts annotating over 15 million signals to improve model accuracy.
Wideer impact on global fisheries
Illegal, unreported, unregulated (IUU) fisheries cost the global economy up to $23 billion a year, accounting for around 20% of the world’s fisheries, according to the Financial Transparency. Particularly affected are African waters, with communities relying heavily on fisheries for nutrition and employment.
Extend AI to wildlife protection
AI2 plans to expand its use of Atlantes beyond maritime applications and integrate it with EarthRanger, a platform that aggregates data from a variety of sources to monitor wildlife. This includes pursuing elephants, rhinoceros and wild dogs to mitigate conflicts in human life. This system is trained to predict elephant behavior and helps reduce conflicts between elephants and farmers.
Future outlook and conservation impact
By utilizing a broad dataset of elephant movements, AI2 aims to actively deal with conflicts in human life. EarthRanger Director Jes Lefcourt highlighted the potential to save elephants’ lives by predicting their movements and preventing conflict with humans.
The infrastructure supporting fishing vessel classification is also applied to predicting elephant behavior, demonstrating the versatility of AI in conservation. This innovative approach could significantly enhance the conservation of natural resources and biodiversity around the world.
Image source: ShutterStock
