OpenAI has alerted US lawmakers about DeepSeek, a Chinese AI startup allegedly attempting to replicate advanced US AI models for its own systems. A memo reviewed by Reuters claims DeepSeek is “free-riding” on the innovations of OpenAI and other American AI labs. The US House Select Committee on Strategic Competition received this warning, highlighting concerns over DeepSeek reportedly bypassing OpenAI’s safeguards to access AI models through masked routes. OpenAI asserts that DeepSeek’s employees are automating the collection of AI-generated responses to train their algorithms. Central to this contention is the method known as “distillation,” which improves new models by leveraging data from established ones, prompting ethical and safety concerns. With DeepSeek gaining recognition for its competitive models, the situation underscores escalating tensions in the global AI landscape, raising issues around model access, copying, and regulation in US-China relations as AI technology evolves.
Source link
