Imazon

PrevisIA is a platform for predicting areas at risk of deforestation in the Brazilian Amazon. PrevisIA was developed through a partnership between Imazon, Microsoft, and Fundo Vale (a socio-environmental investment fund started by the global mining company Vale, headquartered in Brazil). The platform uses an AI algorithm that automatically detects unofficial roads in the forest using satellite images, and then monitors these roads, drawing on evidence that unofficial roads are often risk indicators of deforestation practices including logging, mining, and land grabs. It also draws on data from the past to suggest where deforestation might be most likely to take place in the future.

The focus on predicting future deforestation makes PrevisIA different from other mapping and monitoring projects in the Amazon such as MapBiomas (which researchers from Imazon also worked on) and platforms for monitoring Indigenous territories such as SOMAI and RAISG. Prediction modelling facilitates earlier interventions to protect forest environments. At the same time, the advantages offered by AI monitoring technologies and prediction models cannot be separated from particular constructions of risk and forms of risk governance produced across scientific, state, and commercial discourses.

Find out more in our radio interview with Dr Carlos Souza from Imazon.