Research overview: Zooniverse crowdsourced classification of camera images

Zooniverse hosts citizen-science projects, such as Snapshot Serengeti and its successors, where volunteers worldwide classify camera-trap images, collectively processing millions of photos that would overwhelm research teams. This overview surveys how crowdsourced classification works, with multiple volunteers annotating each image and consensus algorithms aggregating their answers into reliable labels, and how the resulting expertly comparable data both serve ecology and train machine-learning models. It discusses managing volunteer accuracy, handling difficult or empty images, measuring agreement, and engaging the public in conservation science. The overview also notes the synergy between crowdsourcing and artificial intelligence, where humans label training data and verify uncertain machine predictions. By harnessing distributed human effort, Zooniverse made large-scale camera-trap image analysis feasible and seeded the datasets that catalyzed automated wildlife recognition.