Microsoft has publicly confirmed that prohibits its employees from using Deepseekthe most popular Chinese generative artificial intelligence, for reasons related to data security and the risk of propaganda. This was stated by Brad Smith, president of the company, during an appearance before the United States Senate.
“In Microsoft we do not allow our employees to use the Deepseek application,” said Smith, referring to the software that the Chinese company offers both in desktop version and for mobile devices. The manager explained that the decision responds to the concern that the data can be stored on servers located in China, under the legislation of the country, which forces to cooperate with its intelligence agencies. To this is added the possibility that the model responses are conditioned by censorship or the official narrative of the Chinese government.
Deepseek, in fact, declares in its privacy policy that User data is stored on Chinese serverswhich has raised alerts in various organizations and governments. Among them, South Korea, which already banned in February the use of Deepseek applications for privacy reasons, as we collected in these pages. And it is not the only example.
Although Microsoft maintains a version of the DEPEEEK R1 model available in Azure, Smith clarified that it is about a different case, since being open source The model can be downloaded, modified and executed on your own servers without sending data to China. In this sense, the company claims to have carried out an evaluation process to mitigate possible risks, including the modification of the model to eliminate “harmful side effects.”
Despite the hardness of his statements, Smith pointed out that Microsoft does not exclude all competitors in their ecosystem: for example, the perplexity application is available in the Windows store, although not those of Google, including Chrome and Gemini. The exclusion of Deepseek, therefore, is not due to a commercial issue and, according to its criteria, the concrete assessment of the risks it implies.
On the other hand, it is no secret that, although it is not about propaganda as such, The censorship that applies Deepseek On any matter that, it is worth the redundancy, it can question the Chinese regime, it is an easily verifiable fact by any user of the application: it is enough to ask a minimally compromised question so that the double insurance will jump the blockade of the answer. And no, your competition doesn’t work that way.
