Saturday, 28 June 2025
30.1 C
Singapore
32.7 C
Thailand
22.3 C
Indonesia
29.4 C
Philippines

Microsoft bans employees from using the DeepSeek app over data and propaganda concerns

President Brad Smith says Microsoft bans staff from using the DeepSeek app due to data security risks and fears of Chinese propaganda.

If you work at Microsoft, you cannot use the DeepSeek app. The company’s vice chairman and president, Brad Smith, confirmed this during a May 8 U.S. Senate hearing. According to Smith, the ban is due to serious concerns about data security and the risk of Chinese government influence through propaganda.

“At Microsoft, we don’t allow our employees to use the DeepSeek app,” Smith stated, referring to the chatbot service available on desktop and mobile devices. He also mentioned that Microsoft hasn’t added DeepSeek to its app store for the same reasons.

This is the first time Microsoft has spoken publicly about the restriction. Although other companies and even some governments have already taken similar actions, this official stance marks a notable moment in the growing scrutiny of artificial intelligence tools developed in China.

Data stored in China sparks security worries

Smith explained that the ban mainly comes from the risk that user data could be stored in China. DeepSeek’s privacy policy confirms that it stores user data on servers located in China. As a result, that information is governed by Chinese law, which includes requirements to share data with the country’s intelligence services if requested.

Another issue raised was the content produced by DeepSeek. Smith said the app could be influenced by “Chinese propaganda.” DeepSeek is known to censor topics that the Chinese government finds sensitive, making it more likely that users will receive filtered or biased responses when using the app.

Microsoft still uses DeepSeek’s model, under strict control

Even though Microsoft is banning its employees from using the DeepSeek app itself, the company briefly offered DeepSeek’s open-source model—called R1—through its Azure cloud platform earlier this year after the model gained attention for its performance.

Smith clarified that this doesn’t mean Microsoft is endorsing the app. Because DeepSeek is open-source, anyone can run the model on their servers without sending data back to China. This allows companies to maintain some control over privacy and data handling.

However, there are still concerns. Even if the data doesn’t return to Chinese servers, using the model may pose risks. These include the spread of propaganda or the creation of flawed or unsafe code.

During the Senate hearing, Smith said Microsoft had gone into DeepSeek’s R1 model and made changes to reduce what he called “harmful side effects.” He didn’t give many details, instead pointing back to his hearing remarks when media outlets asked.

When Microsoft first made DeepSeek’s R1 model available on Azure, the company said it had undergone “rigorous red teaming and safety evaluations” to ensure its safety.

Not all AI chatbots are banned from Microsoft platforms

It’s worth noting that DeepSeek competes directly with Microsoft’s own Copilot chatbot. However, not all competitors are blocked. The AI search app Perplexity is still available in the Microsoft Store.

That said, a quick search in Microsoft’s store doesn’t show results for any apps by Google, such as the Chrome browser or its Gemini chatbot, suggesting a possible preference for keeping out rival tech giants.

Still, Microsoft’s clear stance on DeepSeek draws a firm line between what the company views as a competitive product and what it sees as a potential threat to its data security and ethical standards.

Hot this week

AWS and DISG launch AI Springboard to help 300 Singapore enterprises adopt AI

AWS and DISG launch AI Springboard to help 300 Singapore enterprises scale AI adoption with funding, training, and technical support.

Lenovo unveils new hybrid AI services and platforms to accelerate enterprise transformation

Lenovo expands its Hybrid AI Advantage with new services, solutions, and platforms to help enterprises scale and operationalise AI.

OIDIRE ODI-XDG10 Portable Baby Bottle Steriliser review: Compact, travel-friendly UV steriliser for modern parents

Compact and stylish, the OIDIRE ODI-XDG10 UV steriliser is a travel-friendly pick for modern parents who want clean bottles without the bulk.

HPE introduces GreenLake Intelligence to simplify hybrid cloud operations

HPE launches GreenLake Intelligence to simplify hybrid IT using AI agents across cloud, storage, networking and sustainability systems.

WhatsApp now lets you use AI to summarise your private messages

WhatsApp now uses Meta AI to summarise your unread private messages, helping you catch up on chats quickly while keeping your privacy intact.

Google adds precise Bluetooth tracking to Pixel Watch 3, but it’s not active yet

Pixel Watch 3 gets new Bluetooth tracking tech called Channel Sounding, which promises precise tracking but still needs full device support.

Meta may buy PlayAI to boost its voice cloning technology

Meta may buy AI voice cloning startup PlayAI to expand lifelike voice features in its apps, smart glasses, and AI assistants.

NVIDIA reveals RTX 5050 entry-level GPU – but is it worth your money?

NVIDIA’s RTX 5050 launches at US$249 with DLSS 3 and Blackwell tech, but better GPU options are only slightly more expensive.

OPPO Reno14 Pro launch offers a limited-time Dyson hairdryer bundle

From June 27 to July 6 in Singapore, get a free Dyson Supersonic hairdryer with OPPO Reno14 Pro or Reno14—this is a limited-time launch offer!

Related Articles

Popular Categories