OpenAI Alleges China’s DeepSeek Stole its Intellectual Property to Train its Own Models

One year on from DeepSeek’s first major release, Washington is probing deeper into the Chinese artificial intelligence firm’s unexpected success.

The U.S. AI firm OpenAI publicly released a memo to the U.S. Congress’s China Select Committee on February 12 alleging that DeepSeek had stolen its intellectual property to fuel its own models. OpenAI noted that DeepSeek continued to steal from ChatGPT over the course of the past year while reportedly preparing to launch a new major model in the coming months.

The allegations showcase China’s reliance on model distillation — training one model using the outputs of another model to lower costs — to circumvent U.S. export controls on advanced artificial intelligence (AI) chips.

OpenAI Provides New Evidence of DeepSeek’s Distillation

The memo alleges that DeepSeek employees use third-party routers and other methods to circumvent OpenAI’s access restrictions, thereby allowing  DeepSeek’s AI models to train on ChatGPT’s systems. OpenAI also claimed that the Chinese firm relied on unauthorized re-sellers of its services to develop more sophisticated distillation methods, along with potential efforts to override ChatGPT’s build-in safety features on chemical and biological weapons development.

The allegations build on OpenAI’s previous claims regarding DeepSeek’s distillation efforts. Shortly after the launch of DeepSeek’s R1 model in January 2025, both OpenAI and Microsoft claimed that it had been partially trained on ChatGPT. This allegation contradicted the firm’s self-reported low training costs and minimal hardware requirements, as rather than relying on its own hardware for training, DeepSeek may have used other, more advanced Western models to speed up its training time. Some AI developers have also claimed that DeepSeek stole from its Gemini series shortly following the June 2025 release of its updated R1 model.

China Prepares to Launch New Models over Lunar New Year

China’s AI sector remains poised to enter a period of heightened growth over the busy Lunar New Year shopping holiday as numerous firms seek to replicate DeepSeek’s unexpected domestic and international success. Chinese state media has lauded the potential flood of releases from Alibaba, ByteDance, and other major firms as a sign of the country’s growth, while major American firms, including AirBnB, have transitioned to using Chinese open-source AI models for customer support and developing internal code bases.

This transition coincides with Beijing’s growing reliance on civilian Chinese AI firms to develop military-specific applications in pursuit of its concept of “intelligentized warfare.” DeepSeek has become an integral part of the People Liberation Army’s (PLA) efforts to streamline its military healthcare system, while other major AI providers, such as Alibaba and Baidu, were listed as Chinese military companies by the Department of Defense in February 2026 for assisting in the PLA’s modernization efforts.  

Washington Should Provide Greater Security Measures

The United States should increase protections against growing Chinese intellectual property theft intended to fuel the PLA’s growth and undermine U.S. economic competitiveness.

While  AI labs in the United States should continue to pursue their own internal security measures to prevent intellectual property theft, Washington can assist them by by bolstering the National Institute of Standards and Technology (NIST). Primarily responsible for AI security, NIST remains understaffed despite the rapid growth of its mission.  Additionally, being subject to unstable congressional funding cycles has left the agency unable to plan effectively to counter Beijing’s industrial espionage campaigns.

Jack Burnham is a senior research analyst in the China Program at the Foundation for Defense of Democracies (FDD). For more analysis from Jack and FDD, please subscribe HERE. Follow Jack on X @JackBurnham802. Follow FDD on X @FDD. FDD is a Washington, DC-based, nonpartisan research institute focusing on national security and foreign policy.