OpenAI has called on the US government to consider banning DeepSeek models, as well as other AI systems developed by organizations linked to the People’s Republic of China (PRC). About this informs TechCrunch.
OpenAI’s proposal notes that DeepSeek’s models, particularly the R1 reasoning model, pose a potential security threat due to China’s data processing laws. OpenAI claims that companies in the PRC are required to provide the government with access to user data upon request, creating risks for international users.
To reduce, as OpenAI notes, “risks to privacy and security, including the risk of intellectual property theft,” the company proposes to ban the use of AI models created in China in “Tier One” countries under the Biden administration’s export restrictions. Those rules already limit the export of high-performance AI chips to China and its allies.
OpenAI does not specify whether its proposal covers the DeepSeek API, open models, or both. At the same time, it is known that many companies, including Microsoft, Perplexity and Amazon, already host open models of DeepSeek on their infrastructure. This may indicate that these models are not necessarily under the direct control of the Chinese government.
OpenAI previously accused DeepSeek of violating its terms of service by “distilling” knowledge from its models. However, recent statements regarding DeepSeek’s ties to the Chinese state indicate an escalation of the conflict between the companies.
OpenAI’s proposal is in line with current U.S. policies aimed at restricting Chinese tech companies’ access to advanced AI technologies. Washington has tightened such restrictions over the past year, citing national security concerns.
