Microsoft has officially prohibited its employees from utilizing those AI application developed by Chinese start up DeepSeek because of the data-security concerns and risk of exposure to the work of the Chinese state-sponsored propaganda. This decision was made public by the Microsoft’s presidency, Mr Brad Smith during the Senate hearing on The race of AI between the U.S and China on the 8th of May, 2025.
Smith commented that the DeepSeek application that exists in the computers and mobile computers stores the users’ data in the servers in China. Such data is available to the government under Chinese law making the issue of privacy critically serious. In addition, the application has indicated cases of censoring regarding matters that the Chinese government finds sensitive – as an effort to steer results towards the state narratives. Among other reasons these caused microsoft to ban the app not only from in the organization but save it from inclusion in app store.
Oddly enough, Microsoft had made DeepSeek’s R1 model available on its Azure cloud service soon after it became famous earlier in the year. However, Smith revealed that this deployment was not the app itself. He noted that Microsoft’s version was “rigorously red team ed and safety evaluated” and removed “harmful side effects” (but, specifics of how these were removed, were not divulged).
This move by Microsoft is one of the growing trends among organizations and governments to make such foreign created AI monitored and at times monitored by the government. It gives the growing role and openness of the data sovereignty and AI systems in the time of siloed regulatory and governance frameworks having been taken to different shores.
Security and integrity of the systems while the AI technologies develop and become part and parcel of every aspect of the daily life and business operations is most important issue. Microsoft’s initiative to go ahead of the rest should be a wakeup call on what’s at risk if the use of AI tools is undertaken without a dike on the way their data handling practices match the organizational values and standards and legal requirements.
