Ai2 and Global Fishing Watch team up on AI ocean monitoring

Ai2 and Global Fishing Watch said Tuesday they are deepening a partnership to build AI agents, satellite analytics and shared models for ocean monitoring and fisheries enforcement. The effort, announced at Climate Week NYC 2026, is designed to help governments and other stakeholders detect suspicious activity faster while keeping human oversight central.

Why it matters: - The partnership aims to help authorities detect, analyze and investigate activity at sea faster, including suspected illegal, unreported and unregulated fishing. - The work could turn large volumes of satellite and ocean data into usable intelligence for governments, civil society and ocean stewards. - The organizations say transparency and human oversight will remain central as AI tools are expanded for ocean governance.

What happened: - Ai2 and Global Fishing Watch announced a strategic partnership at Climate Week NYC 2026. - The collaboration combines Ai2’s AI models and Skylight maritime intelligence platform with Global Fishing Watch’s mapping technology, ocean data and enforcement experience. - The organizations said the goal is to co-develop AI agents and integrated tools for ocean monitoring and enforcement.

The details: - The partnership will integrate the latest satellite data and build new real-time computer vision models. - The organizations plan to explore AI agents that can fuse multiple data sources and surface patterns an analyst might miss. - Global Fishing Watch and Ai2 already collaborate through shared data, vessel registry resources and the Joint Analytical Cell, which supports fisheries enforcement. - In March 2026, Skylight’s Visible Infrared Imaging Radiometer Suite, or VIIRS, went live on the Global Fishing Watch map. - VIIRS adds near-real-time detections of vessel lights at night. - The next phase will co-develop AI agents such as Shippy, Skylight’s AI agent, to support analysis delivery. - Users will be able to ask about a vessel’s history, activity or unusual patterns near a marine protected area or in a region. - The agent will search relevant data and return vessels, events and insights for further investigation. - Ai2 will make OlmoEarth available to Global Fishing Watch for data annotation and model development. - The organizations will jointly develop machine learning models that extract vessel detections and other insights from satellite imagery. - Global Fishing Watch super users will get real-time vessel detections from Ai2’s Skylight platform inside the Global Fishing Watch marine manager portal. - The groups said the tools should support monitoring, control and surveillance of suspected illegal fishing in marine protected areas. - Global Fishing Watch said the collaboration supports its goal of mapping all human activity across the ocean by 2030.

Between the lines: - The partnership reflects a shift from separate AI efforts to a shared detection pipeline and jointly developed models. - The organizations are also positioning the work as a model for responsible AI, with open models, shared tools and impact assessment built in. - The emphasis on human judgment suggests the tools are meant to assist enforcement teams, not replace them. - The focus on open access and transparency also signals a push to make maritime AI more usable for public-interest monitoring.

What's next: - The organizations will keep co-developing AI agents, satellite analytics and machine learning models for maritime detection. - They also plan a series of conversations with environment and technology communities on AI best practices. - The partnership will continue scaling tools that move users from detection to decision more quickly.

The bottom line: - Ai2 and Global Fishing Watch are trying to make ocean surveillance faster, more automated and more transparent without removing humans from the loop.