OpenAI's big claim, ChatGPT was used to create DeepSeek AI
DeepSeek AI: OpenAI said, "We are aware that Chinese companies and others are constantly trying to distill the models of major US AI companies." The company also said that it is working with the US government to protect its models from competitors and potential threats.
Recently, OpenAI claimed that DeepSeek developed its R1 model by distilling its AI models. According to Amar Ujala, OpenAI has said that it has evidence that some users were using the output of its AI models for a competitor, possibly DeepSeek. DeepSeek recently launched its open-source AI model DeepSeek-R1 and hosted it on GitHub and Hugging Face. This model has surpassed OpenAI's o1 AI models in many benchmarks.
According to Amar Ujala, OpenAI alleged that its proprietary AI models were used to train DeepSeek's models. The company also said that it has seen evidence of distillation from several accounts using the OpenAI API. OpenAI and its cloud partner Microsoft investigated the issue and blocked access to these accounts.
OpenAI said, "We are aware that Chinese companies and others are constantly attempting to distill the models of major US AI companies." The company also said that it is working with the US government to protect its models from competitors and potential threats.
AI model distillation is a technique by which knowledge from a large model is transferred to a smaller and more efficient model. This technique helps the smaller model to perform as well as or better than the larger model with fewer computational resources. For example, OpenAI's GPT-4 has about 1.8 trillion parameters, while DeepSeek-R1 has 1.5 billion parameters.
