Microsoft Bans Employees From Using ‘Chinese Propaganda’ Chatbot

Microsoft has banned employees from using DeepSeek — the viral Chinese chatbot it worries is a purveyor of “propaganda” — company President Brad Smith told senators Thursday.

He testified that the app is blocked across all Microsoft devices and stripped from the Windows Store because engineers fear it funnels employee data to Chinese servers and pushes CCP-friendly answers. Smith’s disclosure to the Senate Commerce Committee vaults Microsoft to the front of a widening effort to keep DeepSeek off American — especially government and corporate — devices.

“At Microsoft we don’t allow our employees to use the DeepSeek app,” Smith testified, warning of “data going back to China” and answers shaded by “Chinese propaganda.”

Smith said engineers dissected DeepSeek’s open-source R1 model after it went viral in January and “changed the code … to remove the harmful side effects” after putting it through stress tests designed to catch security gaps and weed out propaganda before listing it on Azure’s cloud marketplace. DeepSeek did not immediately respond to the Daily Caller News Foundation’s request for comment.

The hearing underscored Washington’s growing unease over Beijing-linked AI tools that operate beyond U.S. export controls yet harvest American data. DeepSeek rocketed to the top of Apple’s and Google’s app charts this winter by offering ChatGPT Plus-level performance for free. At the same time, DeepSeek admits it “directly collects, processes and stores your Personal Data in the People’s Republic of China.”

Microsoft’s move goes further than rival OpenAI’s public condemnations of DeepSeek. However, it is consistent with broader trends in the public sector. Lawmakers have already introduced the bipartisan “No DeepSeek on Government Devices Act.”

However, Microsoft’s policy is not absolute. While employees cannot chat with DeepSeek, the stripped-down R1 model remains available on Azure for customers who want to fine-tune it behind their own firewalls — evidence, Smith said, that “open” models can be render safe with enough internal guardrails.

“DeepSeek produced two things,” Smith said. “They have a model that is an open-source model, and they have an application, the DeepSeek app. At Microsoft we don’t allow our employees to use the DeepSeek app, we did not put the DeepSeek app in our app store because of the kinds of concerns that you mentioned — namely data going back to China and the app creating the kinds of content that I think people would say were associated with Chinese propaganda. At the same time, because the model is an open-source model, it was possible for us to go in it, analyze it, and change the code in the model, which we and other people have the permission to do — to remove the harmful side effects.”

DeepSeek’s U.S. ascent has also inflamed intellectual property concerns. Microsoft and OpenAI security teams have investigated whether the startup illicitly “distilled” proprietary GPT models after spotting heavy API scraping last fall.

“We know PRC based companies — and others — are constantly trying to distill the models of leading US AI companies,” an OpenAI spokesperson told Bloomberg News in January. “As the leading building of AI, we engage in countermeasures to protect our IP, including a careful process for which frontier capabilities to include in released models.”

The corporate ban lands as Congress debates how to counter China’s rapid AI advances without hobbling domestic innovators — a balance Smith argued requires expansive American compute power, a lighter federal touch and vigilance against foreign platforms that double as surveillance portals.

For now, every Microsoft-issued employee laptop in Redmond will treat DeepSeek the same way federal IT staff do.