While artificial intelligence (AI) has been positioned as a revolution in the way that people live and work, it has also found an unlikely role in both wildlife and environmental conservation efforts.
Under initiatives by the World Wildlife Fund (WWF), AI technologies have found use in areas like battling illegal wildlife trafficking, understanding the effects of global climate change on the environment, and even potentially finding a way to reverse such effects.
Take for instance their Wildlife Crime Technology Project (WCTP), which kicked off with the installation of Forward-looking infrared (FLIR) cameras for round-the-clock monitoring at Kenya’s Ol Pejeta Conservancy in 2021.
The conservancy, which is home to endangered species such as the African wild dog, has previously been targeted by poachers, making the deployment of these AI-enabled night vision cameras a more concrete form of deterrent against poaching.
In this scenario, AI enters the picture by automating detection, triggering real-time alerts to an operator when the cameras capture humans, wildlife, or vehicles in the places they are deployed.
According to a report from the WWF, similar AI-enabled monitoring systems were also deployed at the Solio Game Reserve in Kenya in 2023, which has not recorded any new poaching incidents since its implementation, according to data available up to 2025.
READ MORE: Protecting marine biodiversity with advanced tech tools
The WWF adds that there have been 11 such deployments across Kenya, specifically targeting rhinoceros sites, resulting in a reduction in such incidents, and in some cases completely curbing them.
AI aids conservation
A similar AI analysis model, known as SpeciesNet, has also found use when it comes to categorising and classifying photographs of animals captured through WWF’s motion-triggered camera traps.
The model has been trained on over 65 million images of camera trap photos sourced from the WWF and other conservation groups, enabling it to detect when an animal is present in an image with 99.4% accuracy. It can also correctly identify species with 94.5% accuracy.
The organisation says it deploys thousands of such camera traps, which can accumulate a staggering amount of data which then needs to be properly identified and sorted.
While doing so manually would be a slow and tedious process that can take months at a time, the WWF says in a separate report that using SpeciesNet can shave months’ worth of manual labour down to just a few minutes.
The model is open-source, meaning that the public – such as researchers, wildlife enthusiasts, and other conservation groups – can freely implement it into their own monitoring systems.
Members of the conservation community can also contribute data to help improve the model’s accuracy through the Wildlife Insights platform. In addition to wildlife identification and image tagging, the platform provides tools for data analysis, including map and graph creation to better identify emerging trends.
This has assisted in conservation efforts in places like the Amazon, where jaguar populations have been threatened by human incursions and habitat loss, giving conservationists a clearer idea of how animals roam and occupy the area.
SpeciesNet is reportedly even capable of identifying individual animals of the same species. This information can then be used to determine the conservation status of various species and accordingly the best intervention methods.
It can even be used to evaluate the effectiveness of initiatives. In one case, the model compiled which species of animals were using a canopy bridge that the WWF installed over forest roads for safer crossings by arboreal (tree-dwelling) species.
Meanwhile, efforts by the Coalition to End Wildlife Trafficking Online (a group made up of major online platforms) have led to the removal or blocking of 63.3 million listings linked to wildlife trafficking as of 2025.
Each member has committed to building AI detection mechanisms, looking at aspects such as the words used in content and image identification for items that violate platforms rules.
The public can also contribute by reporting suspected instances of wildlife trafficking through the coalition’s website. These reports can then be used to further train the respective company’s AI models to automatically detect illegal listings and identify traffickers.
Protecting ecosystems
In environmental conservation efforts, the WWF has deployed both ManglarIA, which looks at supporting mangrove ecosystems, and Forest Foresight, targeting early warning signs of deforestation and forest degradation.
The ManglarIA project in Mexico started in 2023, deploying sensors in the wetlands of the Ría Lagartos. These sensors, which include weather stations, camera traps and drones, are tasked with data collection on the mangrove’s health status.
The data points collected include variables such as temperature, water salinity and flow, and the activity of wildlife in the area, with AI being used to identify patterns and connections between them.
When it comes to Forest Foresight, the project is centred on Peru, particularly the Amazon rainforest, which faces deforestation and forest fires. The WWF says that in 2023 alone, over 132,000 hectares were lost, while another 63,000 hectares were hit by fires.
Forest Foresight is a tool that collates historical satellite imagery, geospatial and socioeconomic data to predict forest loss up to six months in advance through a predictive AI model that analyses past forest loss patterns.
These could be signs, such as expanding roads to facilitate illegal logging or other patterns indicating human activity in the area. Like SpeciesNet, Forest Foresight has also been made open-source and is available for other conservationists to use and build upon.
