IntroductionUse of artificial intelligence (AI) are rapidly increasing in several spheres of healthcare, including electronic health records, diagnostic imaging, radiation therapy planning, robotic surgery, brain implants, wearable technology, chatbots, and others. However, using AI to prescribe drugs has advanced only slowly1,2,3,4. Physician resistance to drug-prescribing AI is perceived to be an important obstacle2,3. Potential variables influencing this resistance, like sex, age, highest education degree, workload, clinical specialty, clinical practicing experience, physician rank, experience with medical AI, hospital tier, and economic development level of a geographic region, are poorly understood. Whether a “doctor-in-the-loop” design—that is, making AI-driven drug prescription conditional on physicians’ discretion—would increase acceptance of drug-prescribing AI is also unknown, as are settings where physicians envision drug-prescribing AI is most likely to be implemented. How physicians’ decision-making processes or values might interact or clash with AI’s is also poorly known5,6.Against the backdrop of progress China has made in medical AI in recent years, attitudes of Chinese physicians towards this issue may provide insights7. In a recent prospective study we conducted, an autonomous conditional drug-prescribing AI model was deployed to assist transplant experts in making complex decisions on a pre-emptive drug intervention to prevent severe acute graft-versus-host disease (GvHD) in persons receiving a human leukocyte antigen (HLA)-haplotype-matched haematopoietic cell transplant8,9. Most physicians and patients agreed to let AI prescribe a drug in this setting, and compliance was high. However, it is unknown whether physicians are receptive to drug-prescribing AI in more common clinical settings.To address this knowledge gap, we designed a large-scale, quantitative survey study that used a stratified sampling strategy to interrogate Chinese physicians regarding their opinions on drug-prescribing AI10. We found most respondents receptive to using drug-prescribing AI and anticipating using it within 5 years. Most respondents prefer conditional to fully autonomous drug-prescribing AI. Respondents suggested initial settings for AI-driven drug prescription include situations where there is little ambiguity in a prescription and situations where prescribing decisions rely on high-complexity clinical data. Male physicians are more optimistic about drug-prescribing AI compared with female physicians, a disparity mediated by differences in medical AI use experience between the sexes.ResultsStratified sampling of participantsSurvey was conducted during 16 May 2025–29 September 2025 in 14 provincial-level administrative divisions (“provinces”) of China. Per capita gross domestic product (GDP) in 2024 was > 15,000 US dollars ($) in 5 of the surveyed provinces (Jiangsu, Fujian, Zhejiang, Tianjin, Inner Mongolia), $10,001 –15,000 in 5 provinces (Anhui, Liaoning, Sichuan, Tibet, Shanxi), and $7001–10,000 in 4 (Qinghai, Henan, Hebei, Guangxi)11.3493 physicians, including 2632 drawn from 32 tier-3 (highest-level) hospitals, 614 from 25 tier-2 hospitals, and 247 from 19 tier-1 (lowest-level) hospitals, participated in the survey (Fig. 1A; “Methods”). The surveyed hospitals are listed in Supplementary Table 1. 271 (7.8%) participants did not complete the survey.Fig. 1: A stratified sample of 2708 physicians.The alternative text for this image may have been generated using AI.Full size imageA Work flow for stratified sampling and data quality control. B Distribution of the 2708 respondents whose returned questionnaires passed data quality checks. “hosp.” and “phys.” stand for hospital(s) and physician(s), respectively.Among the 3222 (92.2% of 3493) participants who completed the survey, median time to completion was 7 min (IQR, 4–12). To ensure data quality, we excluded data from 278 (8.6% of 3222) physicians who completed in