The DeepSeek Ban: Examining Global Concerns Over AI Security and Governance

DeepSeek, a Chinese Large Language Model (LLM), made global headlines in mid-January for introducing the first cost-effective AI model. Unlike Western AI systems, which require billions of dollars and advanced technology, DeepSeek claimed to achieve similar results with significantly lower costs. Its emergence sent shockwaves through financial markets, contributing to declines in American tech stocks. However, as governments and users began engaging with the AI model, concerns over data security, privacy breaches, and geopolitical risks surfaced, prompting multiple nations to restrict or ban its use within sensitive sectors.

The growing opposition to DeepSeek raises critical questions about the safety of AI technologies, especially when linked to national security concerns. Experts warn that the AI model could pose risks such as surveillance, unauthorized data collection, and potential misuse. Given China’s strict cybersecurity laws—which require companies to cooperate with state agencies—many countries fear DeepSeek could serve as a tool for espionage. This geopolitical context has intensified the debate over AI regulation and cybersecurity on a global scale.

National Security Concerns and Government Bans

A key reason for banning DeepSeek in government agencies is the risk of unauthorized data collection and potential national security violations. Security analysts argue that DeepSeek’s underlying mechanisms may enable it to collect and transmit sensitive user data, including IP addresses, conversation logs, and metadata, to Chinese entities. While data collection is common among AI chatbots, concerns stem from where the information is stored and who has access to it. Under China’s cybersecurity laws, businesses must assist government authorities upon request, making foreign governments wary of deploying DeepSeek in official settings.

New York State was among the first to act, banning DeepSeek on all government-controlled devices and networks. Governor Kathy Hochul cited high-confidence reports that the AI chatbot posed risks of unauthorized data transmission. Australia followed suit, imposing similar restrictions, while South Korea’s Ministry of Industry implemented a temporary ban pending a security review. In the United States, lawmakers introduced the “No DeepSeek on Government Devices Act,” seeking to prohibit federal employees from using the AI model.

Beyond data privacy, governments are also concerned about DeepSeek’s vulnerability to adversarial attacks. AI models can be manipulated to generate misleading information, potentially compromising decision-making in public institutions. If AI-generated content is exploited for misinformation or cyberattacks, it could undermine democratic institutions and public trust in governance.

Censorship and Political Bias in AI Models

Another major issue surrounding DeepSeek is its content moderation policies. Reports indicate that the AI chatbot either avoids or manipulates responses to politically sensitive topics. For instance, when asked about the 1989 Tiananmen Square protests, DeepSeek reportedly declined to answer or provided responses aligned with the Chinese government’s narrative. Similarly, discussions on Taiwan’s political status, human rights violations against Uyghurs, or the Hong Kong pro-democracy movement are often censored or framed in favor of the Communist Party of China (CPC).

This level of content control has led critics to accuse DeepSeek of functioning as a propaganda tool. In democratic societies, where access to accurate and unbiased information is a fundamental right, the prospect of AI-generated political bias is deeply troubling. Governments imposing bans on DeepSeek argue that AI systems should prioritize factual accuracy and neutrality rather than reinforce state-sponsored narratives.

U.S. lawmakers have drawn parallels between DeepSeek’s moderation policies and “state-sponsored disinformation,” warning that its use in public institutions could facilitate the spread of misleading or censored information. They emphasize that AI should uphold democratic values, including freedom of speech and access to truth. The controversy over DeepSeek also raises broader ethical questions about AI governance: Who decides what content is censored? And how do global AI models balance accuracy, free expression, and security concerns?

The Global Debate on AI Regulation

The restrictions imposed on DeepSeek reflect a growing urgency for global AI governance. While AI has immense potential to drive progress in healthcare, education, and various other fields, its possible weaponization cannot be ignored. The bans imposed by the United States, Australia, and South Korea reignite the discussion on how AI should be regulated to ensure both security and ethical integrity.

As AI continues to evolve, nations must collaborate to establish transparent regulatory frameworks that balance innovation with security and democratic principles. Without effective oversight, AI technologies risk becoming tools of surveillance, misinformation, or geopolitical influence. The DeepSeek controversy underscores the critical need for international cooperation in shaping the future of AI governance.

Harsh Pandey is a PhD Candidate at the School of International Studies, Jawaharlal Nehru University, New Delhi.