OpenAI Warns Congress That DeepSeek Is Secretly Stealing US AI Data to Train Its Next-Gen Chatbot
Summary
OpenAI warns Congress that Chinese AI rival DeepSeek is secretly stealing outputs from American AI models using obfuscated methods and third-party routers to train its R1 chatbot, raising urgent national security alarms in Washington amid growing concerns over chip export controls and Nvidia's reported technical support to DeepSeek.
Key Points
- OpenAI is warning US lawmakers that Chinese rival DeepSeek is using sophisticated distillation techniques to extract outputs from American AI models, including OpenAI's own, to train its next-generation R1 chatbot.
- DeepSeek employees are reportedly using obfuscated methods, third-party routers, and unauthorized resellers to evade OpenAI's defenses, with the practice posing a serious business and national security threat to US AI companies.
- The situation is intensifying concerns in Washington over chip export controls, with records showing Nvidia provided technical support to DeepSeek, and lawmakers warning that Chinese access to advanced US semiconductors could further accelerate China's AI capabilities.