Archive — US — techcrunch
OpenAI calls DeepSeek ‘state-controlled,’ calls for bans on ‘PRC-produced’models
OpenAI labels Chinese AI lab DeepSeek as 'state-controlled' and urges the U.S. to ban its models, citing security risks tied to PRC regulations. This proposal is part of the Trump Administration's AI initiative. While OpenAI claims DeepSeek's models compromise user data, the exact nature of these models is unclear. DeepSeek, a spin-off of a hedge fund, has attracted scrutiny following a meeting between its founder and Xi Jinping.
Key points
- OpenAI describes DeepSeek as 'state-subsidized' and 'state-controlled'.
- The proposal suggests banning DeepSeek and similar PRC-backed models to mitigate privacy and security risks.
- DeepSeek's R1 model is highlighted as insecure due to compliance with Chinese laws that demand user data provision.
- OpenAI accuses DeepSeek of violating its terms by allegedly using knowledge from OpenAI's models.
- There is no confirmed connection between the Chinese government and DeepSeek, but increasing interest from the PRC is noted.
- DeepSeek's founder recently met with Xi Jinping, raising concerns about government influence.
Why it matters
- The discussion of AI safety and regulations aligns with ongoing global debates about the security implications of AI technologies.
- Concerns over data privacy and potential state-sponsored surveillance have been rising, paralleling similar calls regarding other tech firms.
- The Biden Administration's export rules and focus on AI reflect broader geopolitical tensions around technology and security.
OpenAI's proposal against DeepSeek underscores ongoing tensions between U.S. and Chinese tech sectors and highlights the need for stringent security measures amid growing AI capabilities.