Thus, Trump intends to eliminate competition from the startup that drives the world

On Monday this week, the successful Chinese artificial intelligence app (Deepseek) has been the capital market, when he mimicked a $ 1 trillion from the leading technology companies, of which 600 million Shuler in her son alone, thanks to the proof that you could coach an artificial intelligence model on cheap hardware and low -cost hardware .

The anonymous startup from the Angjo City in East China is praised by senior Silicon Valley, including Sam Altman, CEO of OpenAi, Satya Nadella, Microsoft CEO, super investor Mark Andresen and former CEO of Intel Pat Glasinger, on the innovation he presented, And because of the fact that the use of artificial intelligence will bring it closer to more audiences.

However, in the last day, some warning signs emerged, which will be raised as to the claims of the startups on the meager resources available to him, and fears lest the use of the app holding cyber dangers and possible connection to the Chinese administration. On Tuesday, the White House announced that he opened a test that Depth may endanger the US national security.

White House spokeswoman Carolyn Levitt has confirmed that the National Security Council (NSC) is investigating the situation, and that there is concern among government officials from the implications of the popular app, and its impact on the balance of power in the race for artificial intelligence between the United States and China. Of Trump that “it is a scam to the artificial intelligence industry in the US.”

A ban in the American Navy

The American Navy has also published a warning that prohibits the use of Depth’s technology, indicating security and ethical concerns. The announcement was sent by email to all Navy soldiers last Friday, just days before it became a discretion to the most popular app in Apple’s stores. In the email, the Corps’ Chief Information Officer warned that there was a concern “against the foreign influence of artificial intelligence development” and ordered all soldiers The fleet avoids download, installation or use of the artificial intelligence of Detah, for both work -related tasks and personal tasks.

Not only in the US: in Australia, too, the government warned of Chinese technology, after reports were pushed to do so after the pressure of cyber analysts. Australia’s industrial minister Ed Husik compared Depseau to Tiktok, also owned by a Chinese society, named Bitadans, and said Because “the government should be careful about possible security threats.”

Former Australian Interior Minister Mike Pakhuo, who in 2015 led the decision to block Chinese Wavy’s activities in the country, told Pinnoza Rivio that “every research and development of artificial intelligence has double use: civilian and military, and the significant and effective implementation of China in the field The military may give the army a significant advantage. ”

The foundation for fear of using the popular app rose after some of the myths linked to it turned out to be incorporated. For example, the claim as if it was founded in a tiny $ 5.5 million budget, and came through it to those of OpenAI’s incorrectly: it became clear that the company was linked to a $ 8 billion hedge fund, and both shared a joint CEO. , Were purchased even before American sanctions on her.

Reuters revealed that on the day of the model’s launch, Trump’s day, the company’s CEO Liang Wonpang met with the China Prime Minister in a joint meeting and the Chinese business.

“Be careful about its use”

Is there a real concern about using the app? Gil Masing, a check -in Point spokesman, claims that “in recent days we have identified a vigilant discourse of hackers who are looking for the weak points of Detah, and check how to take advantage of their needs.” However, Masing clarifies, “A similar discourse we have already seen on other artificial intelligence models, including Chatgpt, but until more details about the app, should be careful about using it. For example, it is unknown where its servers are – and despite the fact that it is a Chinese company , This does not necessarily mean that her servers are in China.

Is it a violation of rights?

The fears of the rise of Chinese company also came to Silicon Valley. OpenAi, which announced a cigar, and its sponsor, Microsoft, checks whether the Chinese startup has made a prohibited technology taken from OpenAI, and violated the Terms of Use. It is possible that using ready-made models, or other model data, the Chinese managed to compare to the popular model of the American giant, GPT-O1, but failed to bypass other models, such as Google’s Gemini.

According to Bloomberg, Microsoft security investigators have already identified in the previous quarter that Chinese ruling users have achieved a large amount of data from Openai’s sharing interface to external developers (API). This interface allows to connect Chatgpt’s data and capabilities with other apps, but this can be done with consent to the terms of use -dictated use by OpenAI.

“Microsoft’s accusations focus on the use of a deficit -called technique called ‘dynamic’, or refinement, which includes the use of platform from large language models to improve the performance of smaller models,” says Adv. Yair Geva, head of the high -tech department in Herzog. ” , But there is a concern that Depth has used this technique to train its competing model, which is a violation of Openai’s Terms of Use, and more specifically, the provisions that prohibit ‘rear engineering’, ie observation of the way the code is built to discover the foundations of the software and use its product To develop a competing model. ”

Do US companies have a chance among the legal against Depée? “If they are able to prove that Detana did use the product of OpenAi, it is apparently a blatant violation of its terms of use – because these allow a new model calibration but not creating a competing model.

Attorney Roy Meltzer, head of computing technology in the Ehrlich Group, believes that OpenAi or Microsoft’s chances of winning a faint discount: Use is the lowest in the legal hardware level.

The struggle for ruling is just another stage in the inter -block between China and the United States for controlling artificial intelligence technologies, which also have security implications. In the last five years, the US government has led to the use of the products of a variety of Chinese companies, most notably the chip and Chinese communications company Wawi – according to estimates Also used for the Department’s model training.

“The use of government bodies can be used to use a comma, as they did with most Chinese products, but it will be difficult to prohibit such use of households or private businesses,” says Meltzer. “As for the legislation – if it took more than four years to organize anti -Tiktok legislation, it would be difficult to imagine effective legislation against Depth. – It is very difficult to stop technological development and especially Chinese prosperity in artificial intelligence. ”

Moshe Silberstein, Manager of the NEXT47 Venture Fund, says that “if the United States has blocked Depth, it means that Depsell can not learn and improve from the US use. The model, especially when it may come with spying dangers, so the business impact is very low and temporary. For example, a more powerful model than GPT-O1, but it is true today. ”