With generative AI affecting politics worldwide, researchers face a detection gap," as the biases built into systems mean tools for identifying fake content often work poorly or not at all in the Global South.