Every year the destruction of the rainforests releases more CO2 into the air than all the world’s transportation combined. That’s why Silicon Valley entrepreneur Topher White built an AI system that can be trained to identify the sounds associated with illegal logging. The New York Times caught up with Topher in Indonesia, and the result is a fascinating and well-photographed look at how used cellphones powered by solar panels could change the fight against deforestation.
The challenge of counting forest elephants also offers opportunities for AI applications. Researchers with Cornell’s Elephant Listening Project placed audio recorders in a grid of central African rainforest so dense that it makes spotting the enormous animals by sight nearly impossible. They then turned the audio recordings into spectrograms (pictures of the sound waves) to feed into a neural network for analysis, layer by layer. This NPR story is an entertaining and informative read, and an even better listen.
Listen, or read more