DeepSeek’s overnight success has sparked a nationwide push in China to deploy large language models (LLMs) everywhere from medical institutions to local governments and state-owned enterprises, highlighting Beijing’s eagerness to build on its success in generative neural networks.

Image source: Solen Feyissa / Unsplash

Since Hangzhou-based startup DeepSeek shook up global markets with its R1 AI model last month, China has seen rapid adoption of the technology, fueled by support from companies and government agencies. Every major cloud provider, at least six automakers, local governments, a number of medical institutions, and several state-owned enterprises (SOEs) have begun deploying DeepSeek algorithms. The progress in AI adoption in traditionally conservative government agencies is particularly striking.

«Despite the Chinese Communist Party’s long history of supporting AI, DeepSeek has pushed government agencies and SOEs to adopt large language models. “DeepSeek has changed everything. It has started a nationwide push to promote Chinese AI,” said one SOE technology provider who declined to be named.

The relatively low cost of deploying DeepSeek’s AI models has helped them gain rapid adoption. According to Adina Yakefu, an AI expert at Hugging Face, DeepSeek has “greatly changed the industry landscape” by “lowering the barriers to model development and adoption through its open-source strategy, knowledge-gathering methods, and cost-effective training solutions.”

There are already many examples of successful implementation of DeepSeek AI algorithms. Jilin University Hospital, located in the eastern Chinese city of Changchun, has launched a diagnostic tool that uses DeepSeek’s neural network to create treatment plans, for which the AI ​​algorithm takes data from the hospital’s database, as well as medical recommendations and the effectiveness of different drugs. Jinxin Women’s Hospital in the southwest of the country announced the creation of a tool to track women’s ovulation cycles and create individualized treatment plans for infertility.

Image source: Steve Johnson / Unsplash

One doctor at a public hospital in central China’s Hubei Province said DeepSeek’s AI algorithm has become a sort of referee at their institution, settling disputes when two doctors disagree on the right treatment for a patient. Public hospitals in Chengdu, Hangzhou, and Wuhan have deployed less sophisticated AI algorithms, such as digital nurses that direct patients to the right rooms or explain complex medical terms.

However, several industry insiders cautioned against taking all the AI ​​claims at face value, as some companies are trying to ride investor enthusiasm for DeepSeek without implementing the company’s algorithms. According to a source familiar with the matter, “there’s still a lot of work to do to make these models” capable of performing complex tasks, including medical diagnostics. “They need to be trained on enough medical data to get good results. This will take time and require collaboration with leading AI companies. It’s not something hospitals can build on their own,” the source said.

Even if one is skeptical of some of the claims, experts still point out the willingness to test AI models as a step forward. “The rate of adoption of DeepSeek is incredible. Previously, conservative institutions such as government agencies and hospitals were reluctant to implement generative AI applications, fearing problems if something went wrong,” said one Hangzhou-based AI engineer.

Local governments including the cities of Jinan and Hangzhou have launched citizen chatbots based on DeepSeek models, answering questions from citizens on everything from paying taxes to obtaining a child’s birth certificate. The Futian district of Shenzhen has launched several AI agents based on DeepSeek models, including a tool for law enforcement officers to draft administrative reports.

DeepSeek’s public recognition by Chinese authorities is helping to accelerate the adoption of AI technologies. This month, DeepSeek CEO Liang Wenfeng was invited to meet with Chinese President Xi Jinping, along with business leaders such as BYD CEO Wang Chuanfu, Huawei founder Ren Zhengfei, and Alibaba founder Jack Ma.

Tencent has announced that it has integrated its DeepSeek AI algorithm into the search function of its WeChat app. BYD and Great Wall Motor are among the automakers that have already adopted the startup’s models. Meanwhile, many state-owned companies, including Sinopec, PetroChina, and China Southern Power Grid, have said they have begun implementing DeepSeek’s AI algorithms.

DeepSeek is not profiting financially from the widespread adoption of its technology, according to people familiar with the matter. The company allows its AI models to be downloaded for free and run on public clouds or private servers, making cloud service providers like AliCloud and Huawei Cloud the main beneficiaries of DeepSeek’s surge in popularity.

Leave a Reply

Your email address will not be published. Required fields are marked *