As reported in this week’s New Scientist magazine, artificial intelligence is helping to improve our understanding of the rainforests of Borneo. It’s all down to solar-powered recording devices created to monitor biodiversity, gadgets which are programmed to identify the sounds different creatures make, and therefore track their numbers over time.
It matters because this is a job usually done by humans and as such it’s horribly slow and pretty inefficient, partly because of the terrain and partly because of the area’s sheer size and difficult access. We already have recorders, of course, but they’re battery powered, which means people need to go and change the batteries regularly. Other systems are too expensive, generating impractical amounts of data that gets sent via satellites.
This device is much simpler, a pared down version of a cheap, mass produced mini-computer called the Raspberry PI. In-canopy solar panels provide the power and the data collected is sent via local mobile networks to a remote server. The entire system is leaner, simpler and a whole lot more agile. The machines can easily carry extras like image, audio and video sensors, which collect different types of information including atmospheric data.
The original 12 monitors have already recorded ten thousand hours of audio, which is being used to identify species and watch how their populations change. Another method collects an overall sound map of the forest and looks at how it changes with time, helping to pin down the wider environmental health of the region.
These new AIs will play a vital part in a bigger project designed to track the impact of palm oil plantations in Borneo, hopefully by revealing which tree types protect the greatest amount of bio-diversity. The more we know and understand about the world’s forests, the better we can protect them for future generations.