US Officials Warn: Deepseek Accused of ‘IP Theft’; Microsoft Had Alerted OpenAI

OpenAI, the creator of ChatGPT, claims that Chinese AI startup DeepSeek unlawfully used its proprietary models to train a rival AI system, presenting evidence to support its allegation.

According to the Financial Times, OpenAI identified signs of "distillation"—a technique where developers leverage outputs from larger AI models to enhance smaller ones—which it believes DeepSeek employed. Microsoft’s security researchers first detected suspicious activity last fall, noting that individuals linked to DeepSeek had extracted large volumes of data via OpenAI’s API.

Although distillation is a common AI development method, OpenAI argues that DeepSeek’s actions violated its terms of service, which explicitly prohibit using its outputs to build competing models. In response, OpenAI blocked suspected DeepSeek accounts from accessing its API last year, the Financial Times reported.