OpenAI Pushes for Deepseek Ban, Says Company is 'State Controlled'
OpenAI, the creator of the AI model ChatGPT, describes their competitor DeepSeek as “state-subsidized” and “state-controlled” and submitted a proposal to the U.S. government for a ban on the company and other Chinese AI equipment.
OpenAI Proposes a Ban on DeepSeek
On March 13, OpenAI submitted a proposal to the Office of Science and Technology Policy [1], which is responsible for creating an AI Action Plan to keep America at the forefront of AI development. In the proposal, the company proposed to ban AI models and equipment produced in China that “endanger user privacy and pose security threats, including the risk of intellectual property (IP) theft,” particularly in “Tier 1” nations such as the UK, Canada, and Germany. The firm claims that building systems on top of DeepSeek’s models could be risky as Chinese laws require DeepSeek to provide data that could be used to train models for use by the Chinese Communist Party (CCP). However, the letter doesn’t clearly state whether DeepSeek “models” refer to the R1 model, lab models, and/or API.
OpenAI also suggests other measures it claims will speed up the dominance of U.S. in the global AI landscape. These measures include modernization of government processes to quickly deploy “frontier AI” tools and utilizing the “fair use doctrine” in relation to AI training data.
Concerned about Privacy or Eliminating Competitors?
While the proposal would alleviate privacy concerns, it would also remove one of OpenAI’s competitors, thus making it self-serving. It’s also ironic how OpenAI is concerned about the risk of IP theft, considering the company is facing lawsuits due to copyrighted content being used without permission to train its models [3].
According to TechCrunch, DeepSeek’s open models lack features that would allow the Chinese government to extract user data, as these models are hosted on infrastructures owned by Microsoft, Perplexity, and Amazon. There is also no clear connection between DeepSeek, which was built by a quantitative hedge fund called High-Flyer, and the Chinese government. However, the People’s Republic of China (PRC) has been interested in DeepSeek as its founder, Liang Wenfeng, met with Chinese leader Xi Jinping [2].
The Recent Sucess of DeepSeek
DeepSeek has gained recognition earlier this year as its models delivered impressive performance at a low cost, which allowed China to join the AI race, which was led by OpenAI’s ChatGPT model. Their mobile app quickly amassed over 10 million downloads over the span of two months, which was impressive for a new AI company [4].
OpenAI’s Proposal Doesn’t Aim to Restrict DeepSeek Usage
On March 15, OpenAI spokesperson Liz Bourgeois contacted TechCrunch, saying that they don’t advocate for restricting the usage of models like DeepSeek. However, they proposed amendments to U.S. export rules to allow more countries to get U.S. chips if they promise not to use Chinese technologies in their data centers that could be a potential security risk. The proposal aims to grant other countries and its people access to computing power and AI [2].
References
[1] https://cdn.openai.com/global-affairs/ostp-rfi/ec680b75-d539-4653-b297-8bcf6e5f7686/openai-response-ostp-nsf-rfi-notice-request-for-information-on-the-development-of-an-artificial-intelligence-ai-action-plan.pdf
[2] https://techcrunch.com/2025/03/13/openai-calls-deepseek-state-controlled-calls-for-bans-on-prc-produced-models/?guccounter=1
[3] https://www.pcmag.com/news/canadas-largest-news-organizations-sue-openai
[4] https://blazenwingscorner.space/tech/deepseek-the-chinese-ai-model-that-shocked-the-world/