In a recent policy document, OpenAI characterizes the Chinese AI laboratory DeepSeek as being “subsidized by the state” and “controlled by the state,” urging the U.S. government to contemplate a ban on models developed by this organization, as well as similar entities backed by the People’s Republic of China (PRC).
This proposal serves as a contribution to the Trump Administration’s “AI Action Plan” campaign. It argues that DeepSeek’s systems, such as its R1 “reasoning” model, are vulnerable due to legal obligations under Chinese law that require compliance with user data requests. OpenAI suggests that prohibiting “PRC-produced” models in all countries categorized as “Tier 1” under the Biden Administration’s export policies would mitigate privacy and “security threats,” notably the “prospect of intellectual property theft.”
It remains uncertain whether OpenAI’s mention of “models” refers solely to DeepSeek’s API, its public models, or both. DeepSeek’s open models are designed without features that would enable the Chinese government to extract user data; firms like Microsoft, Perplexity, and Amazon host these models on their platforms.
OpenAI has previously claimed that DeepSeek, which gained notoriety earlier this year, was “distilling” insights from OpenAI’s models in violation of its terms of service. However, the latest allegations—that DeepSeek has affiliations with the PRC and operates under its influence—mark a heightened escalation in OpenAI’s campaign against the Chinese lab.
While there is no definitive evidence linking DeepSeek to the Chinese government, the organization, which is a spin-off from a quantitative hedge fund named High-Flyer, has recently attracted increased scrutiny from the PRC. Just a few weeks ago, DeepSeek’s founder, Liang Wenfeng, had an audience with Chinese leader Xi Jinping.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
