US bans use of China’s DeepSeek by govt agencies

The US government has ordered its agencies to halt the use of artificial intelligence tools developed by Chinese startup DeepSeek, amid a global race for supremacy in the rapidly evolving sector. This directive, issued by the Department of Homeland Security, extends to products created by DeepSeek and those built upon its models by outside entities, mandating their removal from IT systems.

The department retains the authority to instruct civilian organizations, but not military or intelligence services. Defense Secretary's office declined to comment when asked by AFP. DeepSeek's models pose several risks, warns a document accompanying the notice, potentially leading to the inadvertent transfer of sensitive data to China when used by government personnel or contractors.

The servers hosting DeepSeek's operations are situated in China, where the legal framework grants the government considerable reach into private company networks. Attempts to reach DeepSeek for a response have not yet been successful. The AI firm, one of China's leading research institutions, made waves in the industry in January 2025 with the launch of its budget-friendly, high-powered chatbot, challenging perceptions of US technological superiority.

The company unveiled its much-anticipated V4 model in April, followed by an updated, more efficient version of V4 this month. In a display of their intense rivalry, the US has accused several prominent Chinese AI companies, including DeepSeek, of illicitly acquiring advanced model capabilities from American competitors. However, despite their fierce competition, China and the US recently established a communication channel for addressing AI-related incidents, a development that is anticipated to be revisited when their leaders convene at a summit in Shenzhen, China next month.