Strengthening community data

Empowering local communities using artificial intelligence

“…Traditional environmental risk assessment models require a causal link between the risk and the outcome with statistical significance before taking action, which can be very difficult to achieve due to complex relationships between local people and their environments. 29 As a result, citizens collect their own community data (as defined by Carroll et al., 30 such as photographs of smoke emission from a nearby factory) as an alternative to prove their hypotheses. From scientists’ point of view, however, such strong assumption-driven evidentiary collection can lead to biases since the collection, annotation, and analysis of community data are conducted in a manner that strongly favors the assumption.…”