Journalists are pairing satellite and AI to expose illegal mining in the Amazon (www.niemanlab.org)

Investigative journalist Joseph Poliszuk, exiled from Venezuela, has harnessed satellite technology and machine learning to uncover illegal mining activities in the Amazon, showcasing a new model for environmental journalism. With funding from the Pulitzer Center and technical support from Earth Genome, Poliszuk developed a custom machine learning model capable of identifying gold mining sites and related airstrips across vast rainforest areas. His efforts led to the identification of over 3,700 mining locations, some within protected indigenous territories. This innovative approach has empowered journalists to document environmental exploitation from a distance, significantly enhancing the scope and precision of investigative reporting in regions that pose significant access and safety challenges. The initiative marks a pivotal advancement in the AI/ML community as it demonstrates effective applications of geospatial data and AI in journalism. As Poliszuk's model and the Amazon Mining Watch platform continue to evolve, they employ increasingly sophisticated foundation models capable of analyzing various data types beyond satellite imagery. This sets a precedent for similar projects globally, like the upcoming Africa Mining Watch platform, aiming to track mining operations in the Congo Basin. Such advancements are poised to transform investigative journalism, making it more accessible and scalable, while also raising awareness about illegal mining and its impact on vulnerable ecosystems.