Desertification threatens livelihoods and sustainable development in dryland areas, with its risks further intensified by climate change and human activities. However, its extent and severity remain insufficiently quantified across global, regional, and national scales. Here we integrate annual aridity-index-based dryland classifications to indicate potential desertification areas, apply land cover conversion to bare areas to identify desertified areas, and use high-resolution population data to estimate both the number and spatial distribution of populations exposed to desertification at multiple scales. Between 2001 and 2018, the global desertified area increased by 5,387.8 ± 383.5 km² yr⁻¹, while the exposed population grew by 126,857 ± 6,723 people yr⁻¹. Population exposure exhibits pronounced spatial heterogeneity, particularly at the national level. As desertified areas expand, more people, who lack the resources to migrate, become increasingly dependent on degraded land, thereby further accelerating land degradation. This self-reinforcing feedback loop complicates efforts to achieve land degradation neutrality by 2030, as outlined in the United Nations Sustainable Development Goals. By explicitly incorporating hyper-arid areas and linking ecosystem service loss with population exposure, this study develops a cross-scale analytical framework that enhances the assessment of desertification exposure and offers new insights into global desertification risks.