Florida wildlife officials have spent decades trying to control the Burmese python population in the Everglades, an invasive species that has caused severe declines in native mammal populations. In recent years, researchers tested an experimental approach using robotic decoy rabbits designed to lure the snakes into range for capture. The project was a technological failure in its primary mission, but the data it generated accidentally provided researchers with a new understanding of python behavior that has significantly improved removal efforts.

The python invasion began as a human-caused problem in the 1980s and 1990s. Baby Burmese pythons were popular in the exotic pet trade, and many owners were unprepared for how large the animals would become. A mature Burmese python can reach 15 to 20 feet in length and weigh close to 200 pounds. Some owners released their unwanted snakes into the wild near the Everglades. In 1992, Hurricane Andrew destroyed several reptile breeding facilities in South Florida, releasing additional pythons into the surrounding environment.

The Everglades proved highly suitable for the species. The warm climate, abundant water, and prey animals that had no natural defenses against large constrictors allowed the population to expand rapidly. Female pythons can lay up to 100 eggs at a time, and adult snakes have no natural predators in the region.

The ecological damage has been extensive. In areas where pythons are established, raccoon populations have dropped by more than 90 percent. Marsh rabbits and opossums have nearly disappeared from parts of Everglades National Park. Scientists estimate that tens of thousands of pythons now live in the region, though the exact number is unknown because the snakes are difficult to locate in dense swamp vegetation.

Florida has tried multiple removal strategies over the years. The state’s Python Challenge hunting competition in 2013 generated significant media attention but produced only 68 captured snakes. A paid contractor program has removed thousands of pythons annually, but the scale of the territory, roughly 1.5 million acres of difficult wetland, makes comprehensive removal impractical. Specially trained detection dogs proved effective at finding snakes but could not cover enough ground to meaningfully reduce the population.

Researchers later determined that scent is the key to attracting pythons. Early experiments using live marsh rabbits in secure enclosures successfully drew snakes to specific locations, but images of frightened animals surrounded by large pythons caused public criticism, and the program was halted.

Engineers then developed an alternative: robotic decoy rabbits. Each unit cost about $4,000 and was designed to mimic a living marsh rabbit closely. The decoys had synthetic fur, thermal pads that matched the body temperature of a live rabbit, and chemical emitters that released rabbit scent. Some models included motion features such as twitching ears. Each unit also carried motion-activated cameras and transmitters that alerted researchers when something approached.

Researchers deployed 120 decoys across known python habitats. For the first two days, the system appeared to work well. Cameras captured footage of large pythons approaching and striking the decoys, and capture teams removed multiple snakes.

On the third day, the project began to fail. American alligators, which share the same territory and prey species as pythons, started attacking the decoys. Dozens of units were destroyed in a matter of days. At the same time, pythons began avoiding the decoy locations, apparently because they associated the areas with alligator activity. The project was widely considered a failure.

Before the program was shut down, however, a data analyst suggested reviewing the footage collected by the decoys before they were damaged. The cameras had recorded far more than attempted strikes. Timestamps and thermal imagery showed how pythons moved through the swamp. Cross-referencing data from all 120 locations revealed patterns.

The pythons were using specific corridors through the water and sawgrass to travel between hunting grounds. The same routes were used by alligators. The two species had developed behavior patterns that minimized direct encounters. The pythons traveled along these corridors at specific times when alligator activity was lowest. The data effectively mapped the movement patterns of both predator species across the Everglades.

Researchers fed the collected data into a machine learning system. The AI identified correlations between python movements and environmental variables such as temperature, water levels, and time of day. It also identified predictable windows when pythons would be most exposed and easiest to intercept.

Hunters changed their strategy based on this information. Instead of waiting for snakes to come to bait, they positioned themselves along the identified corridors at the times python traffic was highest. The system also helped track mature female pythons, which are the primary drivers of population growth. This led hunters to breeding grounds in remote areas that had not been previously located.

In one month, hunters using the AI-generated intelligence map removed more pythons than the entire previous year of the bounty program. For the first time, removal rates exceeded the snakes’ reproduction rate.

The robotic rabbit decoys failed as bait but succeeded as surveillance platforms. The data they collected continues to inform python management in the Everglades. Officials note that the long-term outcome of the invasion remains uncertain, and whether removal efforts can sustain their current momentum over a large and difficult landscape is still an open question.