Characteristics of media-reported road traffic crashes involving low-level automated vehicles in China, 2015–2025

Background: Low-level automated vehicles (LAVs) have emerged as a new public health challenge, but the epidemiological characteristics of LAV-related crashes remain unknown. Methods: Based on media reports collected by the Automated Road Traffic Crash Data Platform (ARTCDP), we analysed the characteristics of LAV-related crashes in China between 1 January 2015 and 31 August 2025. Results: The ARTCDP captured 4,669 crashes involving LAVs and 324,869 involving other motor vehicles. Compared to other motor vehicles, LAVs were more frequently reported to crash during nighttime (65.3% vs. 29.1%; P < 0.001), on expressways (31.9% vs. 20.2%; P < 0.001), on straight roads without junctions (50.7% vs. 29.4%; P < 0.001), and on rainy days (57.7% vs. 53.6%; P < 0.001). They were primarily reported to crash in economically developed regions, with those in Zhejiang, Guangdong, and Shanghai accounting for 31.6% of all crashes involving LAVs. Furthermore, 20.3% of the LAVs crashes and 16.3% of other motor vehicles crashes were associated with two or more factors. Brake-related issues (48.0% vs. 26.1%; P < 0.01), hazardous road surface condition (55.6% vs. 47.9%; P < 0.01), and distracted driving behaviour (28.7% vs. 9.8%; P < 0.01) more frequently occurred in LAV-involved crashes. Conclusions: Internet-based media reports detected distinct characteristics of road traffic crashes involving LAVs, meriting the attention of policymakers and law enforcement.