The U.S. Department of Homeland Security issued a directive mandating federal civilian agencies to cease the use of all technology products from the Chinese AI lab DeepSeek. This order also requires the removal of any existing implementations of DeepSeek products and services developed using its technology, as first reported by The Register.
The directive broadly targets DeepSeek and its affiliates, as well as third-party products and services that incorporate DeepSeek's technology. This presents a challenge for agencies, as it remains unclear which third-party offerings utilize DeepSeek's models or what constitutes 'developed using.' Services like OpenRouter and Vercel AI Gateway, which route requests to various AI providers, could inadvertently expose agencies to DeepSeek models. Concerns about DeepSeek's AI models stem from allegations that they may collect data on U.S. citizens and transmit it to the Chinese government. Previous actions by the Defense Information Systems Agency, the U.S. Navy, NASA, and the U.S. House of Representatives have already restricted DeepSeek's use within defense and government sectors. Security and privacy flaws were also identified in DeepSeek's mobile applications. Experts note that while banning DeepSeek's direct products is feasible, the broad scope of the order regarding third-party products may be impractical to enforce and verify.
Source: The Register
