OpenAI and Microsoft are investigating whether Chinese AI startup DeepSeek improperly used OpenAI’s technology to develop its new R1 language model. According to reports from Bloomberg and the Financial Times, the companies are probing whether DeepSeek obtained data output from OpenAI’s technology in an unauthorized manner. David Sacks, a venture capitalist and Trump administration member, claims there is “substantial evidence” that DeepSeek used AI distillation techniques to extract knowledge from OpenAI’s models.
The situation has prompted OpenAI to accelerate the launch of its new model, o3-mini, which the company claims was planned before DeepSeek’s emergence. Internal sources at OpenAI reveal ongoing tensions between research and product teams, particularly regarding resource allocation between chat products and advanced reasoning systems. The company faces challenges in managing its dual identity as both a research organization and a commercial enterprise.
The controversy highlights broader industry debates about AI development practices and data usage. OpenAI, currently defending itself against a New York Times lawsuit over training data usage, argues that training AI models on publicly available internet materials constitutes fair use under copyright law. DeepSeek’s success with allegedly fewer computing resources has raised questions about efficient model development and the necessary scale of computing infrastructure in AI advancement.