Researchers studying how economic inequality became a durable feature of societies often point to factors such as agricultural intensification or the rise of the state, but a new study by Dongdong Li and colleagues suggests inequality can have much deeper roots. In their analysis of previously published mortuary data from more than 2,300 burials across 6,000 years of settlement in the Central Plain of China, Li et al. conclude that inequality was shaped by social-relational networks that long preceded the rise of the state. The differentiated social positions within these networks could not be explained by agricultural intensification or the size of the population, they found. But the networks did control an individual’s access to labor, resources, and symbolic capital such as ritual objects. The researchers reconstructed these networks in prehistoric China using isotopic, osteological, and zooarchaeological data from sites dated from 7000 to 1500 BCE. Burial wealth as well as signs of skeletal health and dietary variety were proxies for relational inequalities. Li et al. say that differential access to valued but scarce items initially produced unequal social-relational networks in the region, but that changes in subsistence, the rise of mobile wealth, and escalating violent conflict may have reorganized these networks to concentrate ritual, political, military, and technological control. Early states then “stabilized rather than created these pre-existing disparities by institutionalizing control over resources, violence, and network-based access,” the authors conclude.

In a Columbia University School of Nursing study surveying 1,185 U.S.-based adults about AI use in secure medical portals, 63% of patients said they would not endorse AI-generated medical responses without a clinician's review. The study, “Public Perceptions of Timeliness for AI Use and Human Oversight,” was published in the Journal of the American Medical Association (JAMA) Network Open on October 14, 2026. 

A new technique significantly improves weather forecasting beyond two weeks. Researchers developed a machine-learning model that learns to correct errors in existing forecasting models to make their output more accurate over the two-to-six-week time frame, a key window for emergency planning.

In a study published today in Nature, researchers identified hundreds of new microexons — short gene segments that appear to play a key role in ASD and other neurological conditions — and developed a machine learning approach to predict the impact of ASD-associated genetic variants on the regulation of these segments.