MANILA, Philippines — The governments of the United States, Italy, and Australia have recently banned access to DeepSeek, a Chinese-developed large language model (LLM), on government devices due to concerns over potential data exposure to the People’s Republic of China (PRC). The move reflects increasing global scrutiny of AI technologies developed in China.
However, blocking access to DeepSeek presents unique challenges. As an open-source LLM, it can be run locally on devices, through cloud environments, or integrated into various apps, making complete restrictions difficult to enforce.
Satnam Narang, senior staff research engineer at Tenable, highlighted the complexity of the situation.
“Banning access to the DeepSeek website and mobile applications is straightforward. However, because the DeepSeek LLM itself is open-source, blocking access to it is less straightforward, as it can be run locally on a device, through a cloud environment, and installed via various apps or frameworks like Ollama. It requires some extra leg-work for users to set up, but it can be done,” Narang explained.
While local LLMs reduce the risk of leaking sensitive data by avoiding external servers, concerns remain over the availability of DeepSeek’s website and mobile apps on personal devices, which could still facilitate the sharing of sensitive information.
Another significant concern is the model’s lack of built-in safety features. Narang pointed out that DeepSeek could potentially be used for harmful or nefarious purposes.
As AI models like DeepSeek continue to advance and proliferate, governments worldwide face the challenge of balancing innovation with data security. These restrictions highlight the need for clear policies and advanced monitoring systems to safeguard sensitive information while adapting to the dynamic nature of AI technology.
