Open source, shared planet: Harnessing AI to protect biodiversity

November 13, 2025 

As I attend COP30 this week, I’m grateful to Brazil and its people for reminding us of what’s at stake in the biodiversity crisis. Here, in the cradle of the Amazon rainforest, home to 10% of all known species on Earth, it is clearer than ever why this work matters, and why technology must empower the communities working on the frontlines to protect ecosystems. 

At Microsoft, our ambition has always been to drive innovation that meets the needs of the moment—and biodiversity loss is no exception. Together with partners on the ground, Microsoft is leveraging AI and cloud technology to transform vast planetary datasets into real-time insights, advancing sustainability initiatives at scales once unimaginable.  

Our whitepaper, Investing in Nature for Sustainability, outlines how AI and open-source technology are unlocking new pathways to scale conservation solutions across the globe. For example, Project Guacamaya—a collaboration between Microsoft and several Colombia-based organizations—monitors deforestation in the Amazon region through AI analysis of satellite imagery and wildlife sounds.  

But harnessing AI for good also means ensuring the technology itself is sustainable. That’s why Microsoft is investing in advancing energy efficiency, procuring carbon-free energy, and minimizing the resource use of AI infrastructure—so that the innovation revolutionizing biodiversity efforts isn’t inadvertently harming the planet.  

This month, I'm thrilled to hand the spotlight to my colleague Juan Lavista Ferres, our Corporate Vice President and Chief Data Scientist, who leads the AI for Good Lab at Microsoft. Juan and his team are at the forefront of translating cutting-edge AI research into practical conservation tools that are already making a difference in the field, as he recently described in his TED Talk. 

AI is transforming conservation in ways that were out of reach just a few years ago. The biodiversity crisis spans continents and species, so our tools must be open, adaptable, and collaborative—able to move from one habitat or use case to another with minimal friction. Open source is essential because it lets conservationists tailor models and pipelines locally while contributing improvements back to the global community. 

When we build tools that anyone can adapt and deploy, conservationists can tackle problems we had never anticipated. For example, a solution developed for martens in Washington state can also be used by conservationists to help track jaguars in Brazil, or technology designed for giraffes in Tanzania can be modified to protect tigers in India. Let me tell you about three projects that exemplify this approach: 

🐦 From pilots to platforms: SPARROW’s global network 

Across the world, biodiversity loss is accelerating, but projects like SPARROW (Solar-Powered Acoustic and Remote Recording Observation Watch) show how cross-sector collaboration and AI can help turn the tide. What began as a single pilot is now a growing, connected network in more than six countries—including Colombia, Uruguay, Peru, the US, the UK, and Tanzania—with plans to reach all continents by year’s end. This is where AI is game-changing for wildlife monitoring and protection: models trained to detect and classify species assist conservation researchers in processing huge volumes of data that provide a more robust and timely understanding of the health of vital ecosystems they dedicate themselves to protecting. 

🦒 Pattern recognition for protection: GIRAFFE 

In 1956, a 23-year-old Canadian scientist named Dr. Anne Innis Dagg traveled alone to South Africa to study giraffes in the wild. Her discovery was simple but profound: the pattern of spots on a giraffe is as unique as a fingerprint, allowing scientists to re-identify individual animals over time. Nearly 70 years later, GIRAFFE brings this insight into the digital age through a collaboration between Microsoft and the Wild Nature Institute, using AI to automatically identify giraffes by their distinctive patterns. What makes this project transformative is its open architecture—adaptable to any species with unique markings, from zebras and tigers to whale sharks. While AI alone won't save endangered species, in the hands of passionate scientists and field researchers, it becomes a force multiplier for protecting biodiversity. 

🐟 Removing unseen threats: GhostNetZero.ai  

GhostNetZero.ai began as a local collaboration between WWF Germany and Microsoft’s AI for Good Lab to find and remove ghost nets—abandoned fishing nets that harm marine life while remaining invisible beneath the ocean surface. By combining divers’ firsthand expertise with AI trained on sonar data, the team created a model that can now locate lost nets with remarkable accuracy and efficiency so they can be retrieved. So far, this partnership between WWF Germany and Microsoft technologists has led to the recovery of over 33 tons of ghost nets from the Baltic Sea, with efforts now spreading to key fishing regions such as the Coral Triangle and the North Atlantic.  

AI tools are saving climate and biodiversity researchers countless hours of work. For example, some SPARROW units have logged over 2,000 hours of continuous operation without downtime, processing more than 2 million audio and 1 million video samples. 

This month, we’re featuring a reflection from our partner WWF Germany on the impact of GhostNetZero.ai on marine life worldwide. 

“Ghost nets cause severe damage to marine life: fish, crabs, mammals, and seabirds. We can’t just count on them disappearing; we must act. With the Ghost Net Zero app, we can work with volunteer divers to verify our sonar positions and then feed their reports back into the software. We now want to upscale this process through expanded data that will give us a much broader picture of ghost gear worldwide.” 

— Gabriele Dederer, Research Diver and Ghost Nets Project Manager, WWF Germany  

🔬 GitHub’s Green Software Directory: GitHub Green is a living catalog with over 60,000 climate‑focused open‑source solutions, including AI for Good Lab projects, open datasets, models, and research tools—designed to enable researchers, practitioners, and innovators around the world to tackle some of society’s most pressing challenges. 

💧 FIDO: Since 2023, Microsoft and FIDO have collaborated with water utilities in stressed river basins to reduce water loss from public pipeline networks using AI, helping utilities save resources while improving service reliability. 

🌊 Seaweed‑powered, carbon‑trapping cement: In a recent exploratory study, researchers from the University of Washington and Microsoft used machine learning to develop seaweed‑infused cement that reduces global warming potential without sacrificing strength, an example of how AI can help decarbonize hard‑to‑abate materials. 

Explore our most recent innovation stories at Microsoft Unlocked.

Learn about the work being done to advance sustainability at the AI for Good Lab. 

Read about how Microsoft Azure is promoting sustainable tourism in the Caribbean.