Anonymousabout 2 hours ago
Research groups in Europe and the United States are deploying machine learning models that claim to forecast armed conflicts up to three years in advance, attracting millions in government and foundation funding. But questions about accuracy, data bias against the Global South, accountability gaps, and the absence of consent from affected populations raise serious concerns about whether these tools are ready for operational use — or whether they risk becoming instruments of the power imbalances they claim to address.