How AI has changed the game in wildlife conservation, ocean clean-up

Three weeks ago, a group of scientists published a report revealing that conservationists have been overestimating wildebeest numbers during the annual Serengeti-Mara migration.

Using artificial intelligence tools, the experts concluded that the migrating herds number fewer than 600,000 individuals, approximately half the widely cited estimate of 1.3 million wildebeest, a figure that has remained largely unchanged since the 1970s.

Most aerial surveys in conservation areas have relied on manned aircraft, which can introduce errors since planes usually fly in straight lines over relatively small areas.

Enter the age of AI, revolutionising conservation by replacing rudimentary monitoring systems with digital tools that assess ecosystem health and help deploy resources to critical regions.

Scientists involved in the wildebeest surveys say AI tools minimise double counting by covering much larger areas at once. These methods avoid disturbing wildlife and reduce the risks to human safety associated with aircraft accidents. Already, cloud computing, AI, and Internet of Things imaging tools are being used worldwide to monitor wildlife and forest cover with promising results.

For instance, deep in Costa Rica’s rainforests, a solution known as Clay’s Earth Observation Foundation Model maps and protects crucial, unseen ecosystem elements. This technology safeguards dense biodiversity home to 600 threatened plant species, endangered animals, and around 200 unique trees. “Making satellite imagery available for innovative conservation projects is triggering a wave of practical innovation,” writes Abby Daniell, director of the Multinational Organisations team at Amazon Web Services (AWS). “Data is helping detect changes in Amazon forest cover, monitor coastal waters for illegal fishing, and bring new transparency to agriculture and aquaculture.”

In Kenya, such digital tools could monitor illegal logging in the Mau, Kakamega, or Arabuko Sokoke forests, track wildlife migration in the Maasai Mara National Reserve, and detect encroachment in the vast Tsavo conservation area.

Using AI would enable real-time alerts for deforestation and poaching hotspots, allowing the Kenya Wildlife Service and Kenya Forest Service to respond swiftly, cutting costs associated with extensive ground patrols and reducing risks from ambushes by armed gangs or dangerous animals.

AI tools also offer predictive analytics on droughts, animal migration, and conflict hotspots, helping monitor water levels and reduce human-wildlife conflict in protected areas.

With both public and private sectors seeking to quantify natural resources, digital tools simplify visualisation of biodiversity layers, assess carbon sequestration, and model land-use changes. This ensures communities living near natural resources receive fair compensation for their sustainable stewardship. “AI generates vast amounts of data, unlike manual surveys,” says Nkamunu Patita, co-founder of EarthAcre, an organisation working with indigenous communities to ensure they benefit from their natural capital.

In April 2025, EarthAcre deployed AI tools to scientifically map biodiversity on community land next to Nairobi National Park. The work, led by Andrew Davies of Harvard University’s Davies Lab, helped landowners in the Kitengela Conservation Area receive direct cash transfers, a reward for their long-standing conservation practices rooted in traditional knowledge.

According to Patita, indigenous communities protect 80 per cent of the world’s biodiversity, yet receive little recognition or income due to outdated tracking methods.

Beyond land, AI is helping tackle ocean plastic pollution. The United Nations Environment Programme (Unep) estimates that 19 to 23 million tonnes of plastic leak into aquatic ecosystems annually — a figure expected to triple by 2040 without urgent action. Currently, oceans hold 75 to 199 million tonnes of plastic waste.

Removing waste in remote ocean regions has been a massive challenge for under-resourced conservation groups. To combat this, The Ocean Cleanup, a non-profit, has partnered with Amazon Web Services (AWS) to use AI, machine learning, and cloud computing to remove 90 per cent of floating plastic from the Great Pacific Garbage Patch by 2040.

“Plastic pollution is one of the most pressing environmental challenges. AI can transform ocean data into actionable insights and offer a blueprint for addressing global crises,” says Dr Werner Vogels, chief technology officer at AWS.