Exposed: Scale AI

Scale AI is the dominant data labeling and AI training data company, providing the annotated datasets that power machine learning models for some of the most consequential AI applications in the world including autonomous vehicles, military intelligence systems, and large language models. The company manages a global workforce of over 240,000 data labelers, predominantly in developing countries, who annotate images, text, audio, and video data used to train AI systems. Scale has secured major contracts with the US Department of Defense, US Army, and US Air Force to provide AI-ready data for military applications including target identification, surveillance analysis, and autonomous weapons systems. The company also provides data labeling services for commercial AI companies building self-driving cars, medical AI systems, and generative AI models. Scale data labelers process sensitive content including medical imagery, surveillance footage, personal photographs, and text containing private information, often without comprehensive consent from the individuals whose data appears in the training sets. The working conditions of Scale labor force have drawn criticism, with reports of low wages, algorithmic performance management, and exposure to disturbing content including violence and exploitation without adequate psychological support. Scale AI represents the hidden human infrastructure behind AI systems, where low-paid workers in the Global South process the data that powers AI products generating billions in revenue for Silicon Valley companies.