Is DeepSeek Collecting Too Much?

DeepSeek AI continues to be a major player in the evolving landscape of artificial intelligence, sparking both excitement and concern. Recent developments highlight the complex challenges surrounding this powerful technology, particularly in the areas of data privacy, security, and geopolitical implications.

Recent reports from South Korea's spy agency have raised serious questions about DeepSeek's data collection practices. The agency alleges that DeepSeek "excessively" collects personal data, including keyboard input patterns, and stores this information on Chinese servers, potentially accessible to the Chinese government. This has led to security warnings and restrictions on DeepSeek's use in South Korea, joining a growing list of countries, including Australia and Taiwan, taking similar precautions. 

Furthermore, the South Korean agency has pointed out inconsistencies in DeepSeek's responses to culturally sensitive questions, such as the origin of kimchi, providing different answers in Korean and Chinese. This raises concerns about potential biases and manipulation of information. 

The Open-Source Dilemma

DeepSeek's open-source nature, while fostering innovation and accessibility, also presents challenges for regulation and control. 

According to Sameer Dhanrajani, CEO of AIQRATE and 3AI, DeepSeek stands out due to its rapid reasoning capabilities, cost-effectiveness, and ability to swiftly address cybersecurity threats like phishing and malware.

Unlike other AI-driven security solutions, DeepSeek employs an advanced reinforcement learning technique known as R1, which combines logic, intuition, and reasoning. This makes its response system exceptionally robust and accurate, setting it apart from existing generative AI systems.

One of DeepSeek AI’s major advantages is its industry-agnostic nature. Organizations can build applications on the DeepSeek AI framework across various business functions, enhancing efficiency and productivity. For instance, in HR, DeepSeek can revolutionize talent acquisition, retention, and engagement. 

DeepSeek AI’s reinforcement learning and predictive modeling capabilities also enable effective scenario planning, allowing enterprises to foresee and mitigate cybersecurity risks before they escalate. By leveraging structured and unstructured data, including sources like Wikipedia, DeepSeek minimizes bias and hallucination risks, ensuring reliable AI-driven decision-making.

The DeepSeek AI Ban: National Security Concerns

Despite its technological advancements, DeepSeek AI has been subject to bans on government devices in multiple countries, including the United States (Texas), Italy, and Australia. Satnam Narang, Senior Staff Research Engineer at Tenable, explains that the primary concern stems from fears of sensitive data exposure to the People’s Republic of China (PRC).

Governments have restricted access to DeepSeek’s website and mobile apps, but enforcing a complete ban is challenging due to its open-source nature. The model can be run locally on devices, in cloud environments, and via third-party frameworks like Ollama. This means that while banning DeepSeek AI on official networks is a step toward mitigating risks, individuals can still access it on personal devices, potentially leading to data leaks.

Beyond data security concerns, DeepSeek AI’s lack of robust safety features has raised alarms. The model’s capabilities could be exploited for harmful or malicious activities, prompting increased scrutiny from regulatory bodies worldwide.

As Satnam Narang, Senior Staff Research Engineer at Tenable, points out, banning the website and app is relatively straightforward, but controlling the spread of the open-source LLM itself is much more difficult. This makes it challenging to prevent access and potential misuse, even with government bans in place. 

“Banning access to the DeepSeek website and mobile applications is straightforward. However, because the DeepSeek LLM itself is open-source, blocking access to it is less straightforward, as it can be run locally on a device, it can be run through a cloud environment, and installed via various apps or frameworks like ollama. It requires some extra leg-work for users to set-up, but it can be done. 

 “The local LLMs don’t require access to external servers, minimising the risk of leaking sensitive data. However, banning DeepSeek’s website and apps on government devices is one barrier, because they are still accessible on personal devices, which could be used to share sensitive information. Another concerning aspect to DeepSeek is not just the exposure of sensitive information, but rather the model’s lack of safety features, as it could be used for harmful or nefarious purposes.” Narang adds. 

Global AI Race and India’s Position

Anushree Verma, Director Analyst at Gartner, highlights the broader implications of DeepSeek AI’s rise. China has traditionally been a leader in physical manufacturing, but with DeepSeek, it is now emerging as a key player in AI infrastructure. This shift signals a transition where AI innovation, rather than low-cost physical production, defines global leadership.

For India, which is actively developing its AI ecosystem through initiatives like the National AI Mission, DeepSeek AI’s advancements present both challenges and opportunities. The model’s success underscores the need for India to invest aggressively in AI research and infrastructure to avoid becoming merely a consumer of foreign AI applications. Gartner predicts that by 2028, 33% of enterprise software applications will include agentic AI, a significant leap from less than 1% in 2024. To stay competitive, India must foster AI manufacturing and innovation domestically.

The Need for a Balanced Approach

DeepSeek AI presents a complex dilemma. Its technical capabilities and open-source nature offer significant potential for innovation and accessibility. However, the concerns surrounding data privacy, security, and potential misuse cannot be ignored. The recent reports from South Korea underscore the need for a balanced approach that fosters innovation while addressing the ethical and security challenges posed by this powerful technology.  

Moving Forward

The future of DeepSeek and similar AI technologies will depend on how these challenges are addressed. International cooperation, robust regulations, and ongoing dialogue between governments, industry, and the AI community will be crucial in navigating this complex landscape. The goal is to harness the potential of AI while mitigating the risks, ensuring that this transformative technology benefits humanity as a whole