DeepSeek Is Already Being Applied Widely Across China’s Industries, And Used For Government Surveillance And Propaganda
It’s just three months since the Chinese company DeepSeek released its R1 reasoning model. In that time, its global impact has been dramatic: An article in Fortune magazine described how DeepSeek had “erased Silicon Valley’s AI lead and wiped $1 trillion from U.S. markets.” It also immediately raised serious privacy issues in the EU. The speed at which DeepSeek-R1 was built, and the relatively low cost of doing so, has had another dramatic knock-on effect, shifting interest and investment away from closed development towards open source AI models. As Wired wrote when DeepSeek-R1 appeared:
However DeepSeek’s models were built, they appear to show that a less closed approach to developing AI is gaining momentum. In December, Clem Delangue, the CEO of HuggingFace, a platform that hosts artificial intelligence models, predicted that a Chinese company would take the lead in AI because of the speed of innovation happening in open source models, which China has largely embraced. “This went faster than I thought,” he says.
The open source nature of DeepSeek-R1 means that it is cheaper and easier to use it as the basis of other AI services and products, than to start from scratch. That’s precisely what is happening in China, with the additional twist that many companies are doing so from a patriotic pride that Chinese computer technology has caught up with Silicon Valley. According to Wired:
The narrative that DeepSeek is challenging US dominance in AI has contributed to a growing sense of national pride within China. A central part of the company’s heroic origin story is its development of resource-efficient models, which was seen as a direct response to US policies designed to cut off China’s access to cutting-edge semiconductors. As a result, DeepSeek’s success has fueled a growing belief in China that those measures may eventually fail.
An article on the Rest of the World site quotes Tilly Zhang, a technology analyst at the Beijing-based research firm Gavekal Dragonomics, who points out a key aspect of China’s new-found enthusiasm for DeepSeek:
The government, companies, and investors in China share a common belief that China’s AI opportunity lies more in AI applications — services built on top of AI — rather than solely trying to make models larger to improve performance
The article mention five sectors that are already well advanced with the roll out of DeepSeek to their products. These includes obvious ones like smartphones:
Huawei used the R1 model to upgrade its Siri-like AI assistant, Xiaoyi. The chatbot allows users to pick whether they want their requests to be answered with “automatic deep thinking” or “quick and short responses.” The other big Chinese smartphone makers — Oppo, Honor, Vivo, Xiaomi — have made similar moves by incorporating functions backed by R1, such as text and video creation, and web search.
Carmakers in particular see plenty of scope for DeepSeek AI in their vehicles:
More than 20 Chinese automobile brands have announced plans to embed DeepSeek models, according to local news reports. Much of the adoption appears to be focused on improving existing AI features, including voice control, high-precision mapping, and smartphone-like access to music, web search, and messaging services.
A less obvious applications comes from China’s biggest home appliances company, Midea, which has launched a series of DeepSeek-enhanced air conditioners. Apparently:
The product is an “understanding friend” who can “catch your thoughts accurately,” according to the company’s product launch video.
Healthcare is another major application area:
Nearly 100 hospitals around China have announced they will adopt DeepSeek in their operations, according to a report from Beijing-based online publication Economic Observer News. The applications include supporting diagnostic and treatment processes, analysis of medical imaging, quality control for medical records, and research on new drug uses.
One of the most interesting — and potentially problematic — uses of DeepSeek’s AI technology is by the Chinese government, whose already extreme surveillance systems have been covered extensively here on Techdirt. An article on the China Digital Times site offers a translation of a Chinese post that had compiled more than a dozen local governments’ declarations of DeepSeek’s benefits for monitoring and managing situations online and off. For example, Shandong Province Internet Media Group, a province-level Party media organization, uses DeepSeek’s AI for “public opinion monitoring”:
The group indicated that since it adopted DeepSeek, it has seen great advances in both its efficiency at parsing “public opinion data” from across the internet and its ability to filter out noise, and that it can now more quickly identify potential hazards when monitoring hot topics. In addition, DeepSeek can automatically generate “strategic recommendations for public opinion response” based on its analysis of vast quantities of data, and provide smarter “suggestions for handling public opinion.”
Similarly:
Many local propaganda departments and Party media outlets have said that DeepSeek can automatically generate press releases based on real-time information, helping state-media journalists to quickly compose news reports.
An article by Valentin Weber in the Journal of Democracy predicts that the “next step in the evolution of China’s surveillance state will be to integrate generative-AI models like DeepSeek into urban surveillance infrastructures.” Weber also warns that the rise of AI agents, recently discussed here on Techdirt, is likely to provide the Chinese authorities with even tighter control of their citizens:
The [Chinese Communist Party], with its vast access to the data of China-based companies, could use DeepSeek to enforce laws and intimidate adversaries in myriad ways — for example, deploying AI police agents to cancel a Lunar New Year holiday trip planned by someone required by the state to stay within a geofenced area; or telephoning activists after a protest to warn of the consequences of joining future demonstrations. It could also save police officers’ time. Rather than issuing “invitations to tea” (a euphemism for questioning), AI agents could conduct phone interviews and analyze suspects’ voices and emotional cues for signs of repentance. Police operators would, however, still need to confirm any action taken by AI agents.
This dystopian vision goes beyond even what Elon Musk seems to be striving for, as he rolls out AI technologies with his DOGE project — at least, so far…