TX, NY, VA, and IA State Governments Ban DeepSeek on Government Devices Amid Data Security Concerns
In today’s rapidly evolving technological landscape, artificial intelligence (AI) has emerged as a dominant force in innovation and competition. While industry leaders, such as OpenAI’s ChatGPT, have traditionally dominated the AI landscape, emerging startups are now reshaping the market.[1] Notably, this January, Chinese AI lab DeepSeek introduced an advanced AI model—DeepSeek-R1—that matches or surpasses existing technologies at a fraction of the cost, challenging established players and potentially democratizing access to AI.[2] However, its rapid rise has also sparked concerns surrounding its data practices.[3] Amid growing geopolitical tensions,[4] Texas, New York, Virginia, and Iowa were among the first states to ban DeepSeek from government-issued devices.[5]
Texas Governor Greg Abbott was the first to ban DeepSeek on government-issued devices on January 31, 2025, with the ban applying to all Chinese AI and social media apps.[6] In the press release, he cited cybersecurity concerns, stating that “Texas will not allow the Chinese Communist Party to infiltrate our state’s critical infrastructure through data-harvesting AI and social media apps.”[7] New York Governor Kathy Hochul followed suit on February 10, 2025, implementing a statewide ban on downloading DeepSeek on government devices and networks, also citing cybersecurity and data privacy concerns.[8] The following day, Virginia Governor Glenn Youngkin issued an Executive Order 46 banning DeepSeek on government devices and networks, similarly emphasizing cybersecurity threats.[9] Most recently, on February 19, 2025, Iowa Governor Kim Reynolds issued an executive order banning DeepSeek and other Chinese social media apps from state-owned devices.[10]The press releases and executive orders share a common thread regarding national security and privacy concerns, as each employ language relating to threats and protection.
One of the primary concerns surrounding DeepSeek is its data collection practices, as its privacy policy states that user data is stored on secure servers located in China.[11] This means all conversations, questions, and generated answers are sent to China.[12] While such practice is common among generative AI platforms[13] and it follows naturally that a Chinese company stores its data in China, the concerns stem from the lack of transparency regarding the company’s data privacy initiatives and the potential for data access by the Chinese government.[14]
On the other hand, DeepSeek’s model advancements brought excitement,[15] as it challenged the notion that substantial financial resources are a prerequisite for creating competitive AI technology.[16] DeepSeek made headlines due to its claims that its reasoning capacity is comparable to OpenAI’s o1 models while being 20 to 50 times cheaper, posing a threat to established AI models.[17] This threat was reflected in the stock market, marking a record one-day loss of $593 billion in the market value of Nvidia, the leading manufacturer of AI chips.[18]
Additionally, DeepSeek is open source, meaning its data sets and algorithms are available for anyone to download.[19] This is unlike other AI models available in the market, like Anthropic’s Claude, where its training data and algorithms are not made public, and Meta’s AI models, where anyone can view the codes but its training data is not made public and the application of the models are restricted by licenses.[20] Therefore, anyone can use and build on this model for free.
The data privacy and national security implications of DeepSeek are significant and merit careful consideration. As AI technology continues to advance, it is essential to continue to evaluate privacy and security risks, particularly for technologies with connections to foreign governments. At the same time, these concerns should not overshadow its potential, including its ability to drive innovation, enhance accessibility, and transform industries.
Footnotes
