Adding a political prism to an increasingly contentious rivalry, OpenAI on Thursday called Chinese artificial intelligence (AI) startup DeepSeek “state-sponsored” and “state-controlled” and urged the U.S. government to consider banning its models and others produced in the People’s Republic of China (PRC).
The policy recommendation, submitted as part of the Trump Administration’s AI Action Plan initiative, argues DeepSeek models, including its R1 reasoning model, pose security risks because Chinese laws require companies share user data at the government’s request.
“As with Huawei, there is significant risk in building on top of DeepSeek models in critical infrastructure and other high-risk use cases given the potential that DeepSeek could be compelled by CCP [the Chinese Communist Party] to manipulate its models to cause harm. And because DeepSeek is simultaneously state-subsidized, state-controlled, and freely available, the cost to its users is their privacy and security, as DeepSeek faces requirements under Chinese law to comply with demands for user data and uses it to train more capable systems for the CCP’s use,” Christopher Lehane, OpenAI’s chief global affairs officer, said in a letter Thursday to the Office of Science and Technology Policy.
Separately, Lehane called on the administration to accelerate AI policy and balked at the current regulatory climate under former President Joe Biden. He urged a “voluntary partnership between the federal government and the private sector” instead of “overly burdensome state laws,” paving the way for “freedom to innovate in the national interest.”
Additionally, OpenAI urged a national “export control strategy” for American-developed AI that promotes global adoption of AI systems from the U.S.
But Lehane heaped his strongest scorn at DeepSeek, suggesting the federal government ban AI models developed in the PRC from being used in so-called Tier 1 countries like the U.S. The Biden Administration’s export rules prevent the use of models that could lead to cybersecurity threats, privacy breaches, and intellectual property theft.
OpenAI, however, did not specify whether it was pursuing the prohibition of DeepSeek’s open source models, AI application programming interface (API), or both. DeepSeek is already barred in Italy, Ireland, South Korea, Australia, and large swaths of the U.S. government.
DeepSeek founder Liang Wenfeng’s recent meeting with Chinese President Xi Jinping has stoked fears about the company’s political ties and further inflamed tensions between the U.S. and China in the battle for AI supremacy.
By any measure, OpenAI’s political accusations against DeepSeek on Thursday represent a new escalation of hostilities between the companies.
Previously, the Silicon Valley company accused its Chinese rival of “distilling” knowledge from OpenAI models, which it called a violation of its terms of service.
“The AI industry is at a critical juncture where speed and innovation must be balanced with responsibility and oversight,” Alon Yamin, co-founder and CEO of Copyleaks, said in an email. “While reducing regulatory barriers can accelerate AI advancements and keep the U.S. competitive, removing guardrails entirely poses significant risks, especially in areas like misinformation, cybersecurity, bias, and intellectual property protection.”
