Microsoft Probing If DeepSeek Improperly Used OpenAI Model, Report Says
Key points
- Bloomberg It is reported that Microsoft and OpenAI are investigating whether the group related to Chinese startup Deepseek has used improper data from the American AI company to launch a low -cost AI model sold on Monday.
- Earlier this month, a small part of the cost of AI models such as OPENAI and other American competitors released its super efficient R1 model.
- David Sacks, President of Donald Trump, said in Fox News that there is “a lot of evidence” in Fox News that Deepseek “distilled” knowledge from Openai’s AI model.
Bloomberg Microsoft (MSFT) And OpenAI is investigating whether the group that links with Chinese startup Deepseek link has used the data obtained from the United States incorrectly Artificial Intelligence (AI) The company launched a low -cost AI model, which caused this model Sell on Monday.
InvestOpeDia It has not independently confirmed the investigation. Openai did not immediately reply to the evaluation request. Microsoft refused to comment.
DeepSeek released its super efficiency earlier this month R1 model A small part of the cost of AI models such as Openai and other American competitors. Supported by Microsoft,letter(Googl) He Yuan Platform (YuanTo. This Chinese company said that the performance provided by its R1 model is equivalent to the model of Openai ChatgptEssence
AI Tsar sack said that Deepseek may use the Openai model
this Bloomberg The report was in President Donald Trump “AI and Crypto CZAR“David Sacks) said in Fox that there is” a lot of evidence “that Deepseek uses” distillation | “technology to obtain data from the AI ​​model of OpenAI.
“They can absorb the model from their parents’ models.” Although he did not provide evidence.
nvidia State Tuesday Known as “Deepseek” is “excellent AI progress”, which is called a “perfect example” called the concept of test time expansion.