The Chinese PLA Employs "DeepSeek" at its Military Hospitals

China's People's Liberation Army (PLA) has begun utilising the recently launched Chinese AI tool "DeepSeek" for non-combat support tasks, particularly at military hospitals, to help the doctors create treatment plans in addition to other civilian areas. As per a Chinese media house based in Hong Kong, it stated on 23 March that the PLA hospitals, the People's Armed Police (PAP), and national defence mobilisation institutions are using DeepSeek's open-source large language models (LLMs). According to the Post, the PLA Central Theatre Command's main hospital declared earlier this month that it had approved the "embedded deployment" of DeepSeek's R1-70B LLM, stating that it could offer treatment plan recommendations to assist physicians. Noting that all data was processed and stored on local servers, the hospital also placed a strong emphasis on patient privacy and data security.

Nationwide Deployment

Similar deployments have been observed in other PLA hospitals across the country, such as Beijing's prestigious PLA General Hospital, popularly referred to as "301 Hospital", where extremely sensitive personal data is allegedly housed and prominent Chinese leaders and military officers receive treatment. The PLA, which is making significant investments in modernisation, previously warned its military forces against relying too much on AI, stating that since AI lacks the capacity for self-awareness, it should only be used as a guide rather than a substitute for human judgement in combat. As AI develops, it must continue to be a tool informed by human judgement, according to a January article in the Chinese military's official media house, guaranteeing that responsibility, innovation, and strategic flexibility stay at the centre of military decision-making. Additionally, it said that in order to maximise command efficacy, AI must complement human decision-makers rather than take the place of them.

DeepSeek’s Low Cost Model Attracting Global Eyeballs

China is enthralled with DeepSeek's most recent AI product, which has garnered international notice due to its affordable pricing structure. Furthermore, compared to well-known AI models like ChatGPT, DeepSeek's R1 consumed a fraction of the processing power. Additionally, the US tech industry, which has long defended spending billions of dollars on AI, watched in shock when DeepSeek surpassed ChatGPT as the most popular free software on Apple's App Appstore. Analysts anticipate that the AI models will soon be used in Chinese military decision-making and combat intelligence surveillance. China is encouraging AI integration in a variety of sectors, such as healthcare, manufacturing, and urban development, in addition to treatment programmes in military hospitals. Additionally, certain Chinese government organisations are increasingly using DeepSeek models, notably for anti-corruption initiatives. On its official media account, the Hainan paramilitary force's political work department provided an example of how soldiers used DeepSeek to manage their anxiety and develop an exercise schedule. Before moving into more delicate, high-risk locations, the PLA may attempt to resolve operational and technical issues by first deploying LLMs in non-combat settings, Bresnick told the Post.