OpenAI, the creator of ChatGPT, has raised concerns over what it claims to be the unauthorized use of its proprietary models by Chinese artificial intelligence startup DeepSeek. According to OpenAI, the company has gathered evidence suggesting that DeepSeek utilized OpenAI’s advanced AI models to train its own competing system, in a process referred to as “distillation.”
Distillation is a technique used by developers to enhance the performance of smaller AI models by utilizing outputs from larger, more powerful ones. OpenAI's findings, point to signs that DeepSeek may have used the outputs of OpenAI’s models to refine and improve its own AI capabilities, potentially without proper authorization.
The San Francisco-based AI leader has not yet released detailed evidence to the public but has indicated that its internal analysis shows clear indications of this method being employed. In response, OpenAI has expressed concerns over intellectual property infringement, stating that the use of its proprietary models to train another AI system is not only unethical but also a violation of its terms of service.
While DeepSeek has yet to publicly address the allegations, the accusations highlight ongoing tensions in the rapidly evolving global AI landscape, where the boundaries of intellectual property and model sharing are increasingly coming into focus. OpenAI, known for its groundbreaking work in natural language processing with products like ChatGPT, has long maintained strict guidelines around the use of its models, especially as competition in the AI sector heats up.
As the story develops, the situation is expected to spark further debate about AI ethics, intellectual property, and the competitive dynamics between major global players in the AI space. OpenAI has not ruled out legal action, and the situation may lead to broader discussions on how AI models should be governed and protected in the future.