Microsoft and OpenAI Investigate Potential Unauthorized Data Access by DeepSeek-Linked Group

The AI technology landscape is witnessing intensifying scrutiny as Microsoft Corp. and OpenAI investigate whether a group linked to the Chinese artificial intelligence startup DeepSeek may have accessed OpenAI’s data in an unauthorized manner.

According to sources familiar with the matter, Microsoft’s security researchers detected significant data exfiltration via OpenAI’s application programming interface (API) in the fall. This raises concerns about data security and the potential misuse of proprietary AI models.

Privacy Concerns and Government Scrutiny

DeepSeek, a rapidly rising AI chatbot, has drawn significant global attention due to privacy and data security concerns.

Australia’s Science Minister, Ed Husic, became the first Western government official to openly raise concerns about the app, emphasizing the need for careful evaluation of its data and privacy management practices.

Given past allegations against Chinese tech firms regarding links to the Chinese state, there is growing apprehension over the possibility of DeepSeek harvesting user data for intelligence purposes.

U.S. President Donald Trump acknowledged DeepSeek’s disruptive potential, calling it a “wake-up call,” but he stopped short of labeling it a national security threat. Instead, he highlighted its potential benefits in reducing AI costs.

Nevertheless, governments remain cautious, and privacy advocates are calling for deeper investigations into how DeepSeek manages user data.

DeepSeek’s Soaring Popularity and Data Collection Practices

Despite these concerns, DeepSeek has surged in popularity, topping app store charts in both the U.S. and the UK.

According to market analysts Sensor Tower, the app has been downloaded three million times since launch, with 80% of these downloads occurring in just the past week—three times the rate of competitors like Perplexity.

DeepSeek’s privacy policy reveals that it collects and stores extensive user data on secure servers in China. This includes:

1. Email addresses, phone numbers, and dates of birth provided during account creation.

2. All user input, including text and audio interactions, as well as chat histories.

3. Technical details such as device models, operating systems, IP addresses, and even keystroke patterns.

The company claims that the data enhances the service's safety, security, and stability. However, the policy also states that DeepSeek shares this information with service providers, advertising partners, and corporate affiliates and retains data for as long as necessary.

Read Also: Understanding DeepSeek R1 and Its Approach to AI Reasoning

Expert Concerns and Regulatory Responses

Privacy experts have expressed alarm over DeepSeek’s data handling practices. Lauren Hendry Parsons, a digital privacy advocate at ExpressVPN, highlighted a particularly concerning clause in the app’s privacy policy, which states that user data may be used "to help match you and your actions outside of the service."

She warned that this should serve as an immediate red flag for users worried about their privacy.

While DeepSeek’s data collection practices may appear excessive, analysts point out that similar policies are in place for AI services such as ChatGPT and Google’s Gemini. 

However, security experts caution that any AI model with a web or app interface inherently exposes user prompts and responses to the developers.

Dr. Richard Whittle from the University of Salford advised users to be cautious, especially given the hype surrounding DeepSeek. Meanwhile, the UK’s Information Commissioner’s Office (ICO) urged users to stay informed about their data rights and AI’s data collection practices.

The agency emphasized the importance of transparency among AI developers and affirmed its commitment to enforcing data protection regulations.

The Role of Chinese AI and Huawei’s Involvement

DeepSeek’s meteoric rise is also tied to its significant technological advancements. The app operates on a Huawei-developed AI chip, the Ascend 910C, which powers inference tasks—processing and generating responses based on trained AI models. 

While DeepSeek’s large-language model (LLM), DeepSeek R1, was initially trained using Nvidia’s H100 GPUs, it now relies on the Huawei chip for real-time operations.

This revelation, brought to light by tech analyst @Dorialexander, suggests that the Chinese AI sector is moving toward self-sufficiency in AI hardware.

However, the Ascend 910C’s relatively lower performance compared to leading Western alternatives limits its use for training AI models.

To address this, Huawei is reportedly developing the more powerful 920C chip to rival Nvidia’s Blackwell B200, a leading AI chipset in the industry.

Read Also: DeepSeek AI vs. Scammers: The $48M Crypto Fraud You Need to Know About!

The Future of AI Competition and Data Privacy

The escalating competition in AI development underscores broader concerns about data privacy, security, and geopolitical implications.

With DeepSeek disrupting the AI landscape and posing new challenges to industry leaders like OpenAI, regulators and consumers must navigate emerging AI technologies' risks and rewards.

Governments worldwide are increasingly prioritizing AI governance, ensuring that data privacy measures are enforced while fostering innovation.

As investigations into DeepSeek’s data practices continue, the outcome may shape the global AI regulatory landscape and influence how future AI applications handle sensitive user information.

For now, users are advised to exercise caution and remain informed about the terms and conditions governing their personal data in an era where AI-driven services are rapidly transforming digital interactions.

FAQ about DeepSeek AI

1. What is the main concern regarding DeepSeek and OpenAI?
Microsoft and OpenAI are investigating whether a group linked to the Chinese AI startup DeepSeek accessed OpenAI’s data without authorization. This raises concerns about data security and the potential misuse of proprietary AI models.

2. How was the unauthorized data access detected?
Microsoft’s security researchers detected significant data exfiltration through OpenAI’s API in the fall. The specifics of how the access occurred are still under investigation.

3. Why is DeepSeek under scrutiny?
DeepSeek has drawn attention due to concerns about privacy and data security, particularly regarding its data collection and storage practices. The company’s links to China have also contributed to apprehensions about potential government involvement.

4. Has any government official raised concerns about DeepSeek?
Yes, Australia’s Science Minister, Ed Husic, was the first Western government official to raise concerns about DeepSeek’s data and privacy management practices. Other governments are also closely monitoring the situation.

5. Where does DeepSeek store user data?
The company stores user data on secure servers in China. It also shares this data with service providers, advertising partners, and corporate affiliates.

6. How does DeepSeek justify its data collection?
DeepSeek claims that data collection enhances the safety, security, and stability of its services. However, privacy advocates warn that the policy allows extensive data retention and sharing practices.

7. How does DeepSeek compare to other AI services like ChatGPT and Gemini?
While DeepSeek’s data collection practices are considered extensive, similar policies exist for other AI services, such as ChatGPT and Google’s Gemini. However, the key concern is the potential for government access and misuse of collected data.