OpenAI has publicly accused DeepSeek of illegally distilling its models — essentially using ChatGPT outputs to train DeepSeek at a fraction of the cost. The White House AI czar David Sacks said there is substantial evidence of this. DeepSeek admits its models were influenced by OpenAI outputs but denies wrongdoing. Harvard Law professors argue OpenAI has a weak legal case since its own models were trained on copyrighted data without permission. The irony is hard to miss. Did DeepSeek steal from OpenAI or is this just how AI development works?
