LOGO

Microsoft Bans DeepSeek App for Employees - Details

May 8, 2025
Microsoft Bans DeepSeek App for Employees - Details

Microsoft Prohibits Employee Use of DeepSeek

Microsoft employees are restricted from utilizing DeepSeek due to concerns regarding data security and the potential for propaganda, as stated by Microsoft Vice Chairman and President Brad Smith during a recent Senate hearing.

Smith confirmed that the DeepSeek application service, accessible on both desktop and mobile platforms, is not permitted for use by personnel within Microsoft.

Concerns Regarding Data Storage and Influence

The decision to block DeepSeek extends to its inclusion in Microsoft’s app store, stemming from apprehensions about data storage locations and potential external influence.

This marks the first public announcement by Microsoft regarding a prohibition of this nature, despite similar restrictions being implemented by various organizations and governments.

Smith articulated that the primary risk lies in the possibility of user data being stored on servers located in China.

Furthermore, there are concerns that responses generated by DeepSeek could be subject to manipulation through “Chinese propaganda.”

DeepSeek’s Privacy Policy and Censorship

DeepSeek’s privacy policy explicitly states that user data is stored on Chinese servers.

This subjects the data to Chinese law, which necessitates cooperation with Chinese intelligence agencies.

The platform also implements substantial censorship of topics deemed sensitive by the Chinese government.

Microsoft’s Dual Approach

Interestingly, despite these criticisms, Microsoft made DeepSeek’s R1 model available on its Azure cloud service shortly after the model gained widespread attention.

However, this differs significantly from offering the DeepSeek chatbot application itself.

As DeepSeek is open source, the model can be downloaded, stored on independent servers, and offered to clients without transmitting data back to China.

Remaining Risks

Even with this approach, potential risks remain, including the dissemination of propaganda and the generation of insecure code.

Modifications to the DeepSeek Model

During the Senate hearing, Smith revealed that Microsoft had successfully accessed and “changed” DeepSeek’s AI model to mitigate “harmful side effects.”

Specific details regarding these modifications were not disclosed, with TechCrunch being directed to Smith’s original statements.

Microsoft previously stated that DeepSeek underwent “rigorous red teaming and safety evaluations” prior to its release on Azure.

Competitive Landscape

It is worth noting that DeepSeek’s app presents direct competition to Microsoft’s own Copilot internet search chat application.

However, Microsoft does not universally ban competing chat applications from its Windows app store.

For example, Perplexity is currently available in the Windows app store.

Conversely, applications developed by Microsoft’s primary competitor, Google – including the Chrome browser and Gemini chatbot – were not found during a search of the webstore.

#Microsoft#DeepSeek#app ban#AI#employees#technology