White House accuses DeepSeek of copying OpenAI technologies
A major scandal erupted around Chinese company DeepSeek following a statement by David Sacks, White House’s chief AI advisor. According to him, there is “substantial evidence” that the Chinese developer used OpenAI technologies to create their new model.
In a Fox News interview, Sacks revealed a technique known as distillation, where one AI model uses another model’s outputs for training and developing similar capabilities. “There is substantial evidence that DeepSeek applied knowledge distillation from OpenAI models, and I think OpenAI is not at all happy about this,” said the advisor, without disclosing specific details.
The situation escalated after DeepSeek presented its new open-source model R1, capable of imitating human thinking. The company claimed that R1 not only matches leading American developments across several industry metrics but surpasses them, while development costs were significantly lower.
Notably, OpenAI CEO Sam Altman has already begun an internal investigation to determine whether DeepSeek’s successes are indeed related to OpenAI models’ distillation rather than being the result of an independent technological breakthrough.
Despite criticism, Sacks acknowledged DeepSeek’s achievement in creating a more efficient model without using many advanced GPUs. However, he warned that the buzz around the Chinese company would push American AI developers to take measures preventing distillation to slow down the emergence of so-called “copying” models.
OpenAI is currently refraining from official comments about the situation, which only fuels the tech community’s interest in the unfolding confrontation between American and Chinese artificial intelligence developers.
