Business and technology journalist
Openai, a manufacturer of CHATGPT, complained that rivals, including China, are trying to quickly develop their AI tools.
As a world leader of AI, the status of Openai and other US companies This week was rapidly damaged The Chinese app, which can imitate the performance of CHATGPT with the sudden appearance of DeepSeek, is clearly part of the cost.
Bloomberg has reported that Microsoft is investigating whether the data (major investors) belonging to Openai was used unauthorized.
The BBC contacted Microsoft and Deepseek and presented opinions.
Openai’s concerns were reversed by David Sacks, the recently appointed White House “AI and Crypto Czar”.
In Fox News, he suggested that DeepSeek would have improved the process called knowledge distillation using the model developed by Openai.
Sacks said, “There is practical evidence that DEEPSEEK distilled knowledge in the model of Openai.
“I think that one of the months to see is a major AI company that takes measures to prevent distillation. This will slow down some of these copy cat models.”
The United States has already taken measures to protect the development of AI, and China has suggested a rules for cutting China in high -end chips and controlling investment in the United States under the name of national security.
Trump, a candidate for Howard Rutnick, has raised the prospects for US measures to share concerns about theft and protect the US AI company at his confirmation hearing on Thursday.
Lutnick said, “What this showed is that export control supported by tariffs is like a stupid model.
In the statement, Openai said, “China and other companies are trying to constantly separate the model that leads the United States.”
“When we go forward … It is very important to work closely with the US government to best protect the most competent models.”
‘Fraud’ claim
Naomi Haefner, a technical management assistant at the University of St. Gallen, said that distillation problems could be made for some of the costs of deepSeek.
“I’m not sure if DeepSeek actually trained the model from the beginning,” she said.
“Openai believes that Deepseek would have misused a large amount of data from them.
“If this is true, the claim for model training is very cheap and deceptive. We don’t know if we are actually possible until someone replicates the training approach.”
Crystal van Oosterom, an AI venture partner of Openocean, agreed that “Deepseek has clearly established a publicly available study of major US and European institutions and companies.”
But it’s not clear how the idea of ”construction” is a problem for others’ work.
This is especially true in AI, where accusations of ignoring intellectual property are frequent. Leveling from Major US AI company.
Security and ethics
According to the White House reporter, Karoline Leavitt, US officials are also considering the influence of national security on the emergence of Deepseek.
“I talked with this morning (National Security Council) and spoke. They reaffirmed the US president’s remarks a day ago. Call the US technology industry.
The announcement was announced after the US Navy banned the use of Deepseek’s app due to “potential security and ethical problems.”
According to CNBC, the US Navy sent an email to employees not to use the DeepSeek app because of the potential security and ethical concerns related to the origin and use of the model.
The Navy did not immediately respond to the BBC NEWS request.
Data safety experts should be careful to the user if they collect a large amount of personal data from the user and store them on a Chinese server.
Meanwhile, Deepseek said that it is the target of cyber attacks. On Monday, Software’s “massive malicious attack” will temporarily limit registration.
The banner displayed on the company’s website says that the registration can be busy as a result of the attack.
Fan King’s additional report
