Chinese competitors use their work for artificial intelligence applications

Chatgpt, Openai, complained that competitors, including China, used its work to make rapid progress in developing artificial intelligence tools (AI).

The case of Openai – and other American companies – such as world leaders in artificial intelligence It was greatly undermined this week By the sudden appearance of Deepseek, a Chinese app that can simulate Chatgpt performance, apparently in a small part of the cost.

Bloomberg stated that Microsoft has realized whether the data belonging to Openai – a major investor in it – has been used in an unauthorized manner.

BBC called Microsoft and Deepseek to comment.

The recently appointed White House “AI and Crypto CZAR”, David Sacks.

Speaking on Fox News, he suggested that Deepseek had used the models developed by Openai to improve, a process called distillation of knowledge.

Mr. Sachs said: “There is fundamental evidence that what Dipsik did here is that they distorted the knowledge of Openai.”

“I think one of the things that you will see in the next few months is the leading artificial intelligence companies that take steps to try to prevent distillation … This will definitely slow down some of these counterfeit models.”

In a statement, Openai said that Chinese companies and other companies “are constantly trying to evaluate the driving models of the American companies.”

He added: “With our progress … it is very important that we are working closely with the United States government to protect the most capable models.”

The issue of distillation could throw the idea that Dibsic has created its product for a simple part of the cost, said Naomi Hefner, an assistant professor of technology management at St. Galen University in Switzerland.

And she said: “It is not clear whether Dibsic has trained his models from scratch.”

Openai stated that they believe that Deepseek may have been abuse of large amounts of data from them.

“If this is the case, the claims related to training the model at a very lowly deceptive price. Until someone repeats the training approach, we will not know with certainty whether this cost -effective training is really possible.”

Crystal Van Oosterom, AI Venture partner in OpenOCEAN, agreed that “Deepseek has been clearly built on research available to the public from major American and European institutions and companies.”

However, it is not clear how clear the idea of ​​”building on” the work of others.

This is especially true in artificial intelligence, as the accusation of unauthorized intellectual property rights was repeatedly Played in the leading American AI companies.

US officials are also studying the effects of national security for the emergence of Deepsik, according to White House journalist Caroline Levitte.

“I spoke with the National Security Council this morning,” said Ms. Levitt, who also re -re -reported US President Donald Trump this morning.

This announcement comes after the US Navy prevented its members from using Deepseek applications due to “potential security and moral fears.”

According to CNBC, the US Navy sent an e -mail to its employees warning them against using Deepseek due to “possible and moral security associated with the origin of the model and its use.”

The navy did not immediately respond to a request for comment from the BBC News.

Data safety experts have warned users not to beware of this tool, given that they collect large quantities of personal data and store them in servers in China.

Meanwhile, Dibsic says it was the goal of electronic attacks. On Monday, she said it would be temporarily limited to records due to “large -scale harmful attacks” on her programs.

A banner appears on the company's website says that registration may be busy as a result of the attacks.

Additional reports from the fan Wang