US Considering Restrictions on Chinese AI Firm DeepSeek Over Security Concerns

The White House is evaluating various measures to limit the presence of Chinese artificial intelligence company DeepSeek in the United States due to national security concerns, according to The Wall Street Journal. Officials are particularly focused on how DeepSeek manages user data, which the company acknowledges is stored on servers in China.

The administration of President Donald Trump is expected to introduce a rule preventing the download of DeepSeek’s chatbot application on government-issued devices, The Wall Street Journal reports, citing sources familiar with the matter. Authorities are concerned that DeepSeek has not sufficiently clarified its data collection policies, including how the information is used and who has access to it.

Beyond restricting government use, officials are also considering two additional actions: a full ban on the DeepSeek app in US app stores and regulations limiting American cloud providers from offering DeepSeek’s AI models. However, discussions on these measures remain at an early stage, sources told The Wall Street Journal.

DeepSeek caused a stir in both Silicon Valley and Washington earlier this year when it introduced a powerful AI model developed at significantly lower costs than its American competitors. For a brief period in January, DeepSeek’s chatbot was the world’s most downloaded app, though it later fell to seventh place in February, according to data from Sensor Tower.

Several countries have already imposed restrictions on DeepSeek. The company voluntarily removed its app from Italian app stores, while South Korea halted new downloads due to unresolved concerns about its privacy policy. Additionally, Australia, Canada, and Taiwan have prohibited the use of DeepSeek on government devices.

Although no overarching ban exists within the US government, specific agencies, including the Navy and NASA, have already barred the app over security and privacy risks. Some states, such as New York and Texas, have also taken steps to curb its use.

A complete ban on DeepSeek for the general public presents greater difficulties than restricting it from government devices. The company’s AI models are available as open-source software, meaning they can be freely downloaded and replicated. US cloud providers currently offer DeepSeek’s models as part of their services, and limiting this access could face pushback from businesses and investors advocating for open technology sharing.

“US officials considering technology controls are dealing with new territory here on what to do with open-sourced models,” Paul Triolo, a partner at DGA-Albright Stonebridge in Washington, D.C., told The Wall Street Journal.

Companies using open-source AI models can operate them on their own infrastructure, allowing them to maintain control over data security while addressing national security concerns, Triolo further explained.

Silicon Valley AI search firm Perplexity, which provides DeepSeek’s open-source model to customers, stated that the model runs on secure US-based data centers independently of DeepSeek’s main website.

When asked about DeepSeek at a press conference, Chinese Foreign Minister Wang Yi criticized US actions, stating that efforts to cut China off from technological advancements would backfire. “Where there are blockades, there are breakthroughs,” Wang said, according to The Wall Street Journal.

The restrictions under consideration by the Trump administration are separate from a bipartisan bill introduced in Congress last month that seeks to prohibit the app on federal government devices.

Leave a Reply

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