OpenAI is suspicious of the Chinese AI competitor that has become a household name overnight.
OpenAI suspects that DeepSeek, a low-cost AI model maker, improperly used the company’s data to train its AI models. It is also investigating that.
Microsoft, OpenAI’s main investor, is also investigating whether DeepSeek used OpenAI’s technology to train its AI. David Sedgwick, the new White House AI and Crypto chief, has said that he believes DeepSeek is using OpenAI’s models to improve its performance.
In Sedgwick’s terms, DeepSeek is using “knowledge fusion,” which is a way of “synthesizing” itself. Sedgwick says he has evidence that DeepSeek is filtering out knowledge from OpenAI models.
OpenAI is investigating DeepSeek after these allegations emerged. DeepSeek has also shown AI gurus how to do incredible things with its inexpensive but powerful LLMs. OpenAI CEO Sam Ottman himself praised DeepSeek as a worthy model on the X.
On the other hand, after releasing a high-end AI model at a low price, US stocks, including those of leading US chipmaker Nvidia,
fell sharply. If DeepSeek really borrowed technology from OpenAI, it would be a big problem.