OpenAI Accuses DeepSeek of Distilling U.S. Models, Reports Bloomberg

OpenAI alleges that DeepSeek is distilling results from leading U.S. AI models to train its R1 chatbot, according to a Bloomberg report citing an internal memo. The accusation was made to U.S. lawmakers. DeepSeek has not issued a public response. The report dates to February 12, 2026.