OpenAi accuses Chinese rivals of copying their AI models

American says Chinese companies are using their AI models to develop rival tools, raising concerns about safety and intellectual property in the artificial intelligence sector

OpenAi, ChatgPT’s creator, expressed concern for rivals, including Chinese companies, which would be using their research to accelerate the development of their own artificial intelligence tools (AI).

OpenAi’s position – and other US companies – such as Global AI leaders has been shaken this week with the sudden emergence of DeepSeek, a Chinese app that seems to replicate chatgpt performance at a significant cost.

Bloomberg said Microsoft is investigating if data from OpenAi – of which it is one of the main investors – were used unauthorized. The BBC contacted Microsoft and DeepSeek for comments.

OpenAi’s concerns were reinforced by David Sacks, recently appointed “AI and Cryptocurrency Csar” of the White House. In an interview with Fox News, he suggested that DeepSeek may have used OpenAi models to improve its performance, a process known as knowledge distillation.

“There is substantial evidence that Deepseik has unleashed the knowledge of OpenAi models,” Sacks said. “In the coming months, we will see leading companies to take action to prevent distillation … This would certainly slow some of these copied models.”

The US has already taken steps to protect its AI advances, with rules aimed at cutting China’s access to advanced chips and directing investments to the country in the name of national security.

In his confirmation hearing on Thursday, Howard Lutnick, nominated from Trump to Secretary of Commerce, also expressed concern about intellectual property theft and suggested that the US can take more actions to protect their AI companies.

“What this has shown is that our export controls, without tariffs, are like a Whack-a-Mole game,” said Lutnick.

In a statement, OpenAi has stated that Chinese companies and others “are constantly trying to distill the models from major US AI companies.”

“As we advance … It’s crucial to work closely with the US government to protect the most capable models,” the company added.

“It’s unclear if DeepSeek really trained its models from scratch,” she said. “OpenAi has stated that it believes that DeepSeek may have improperly appropriated large amounts of data. If true, low -cost training allegations are misleading. Until someone replicates the training approach, we will not know for sure if it is really possible. ”

Crystal van Oosterom, OpenOcean AI partner, has agreed that “DeepSeek was clearly based on public research from large American and European institutions and companies.” However, it is unclear how problematic is the idea of ​​“building on” the work of others, especially in the AI ​​area, where accusations of disrespect for intellectual property rights are frequent.

Security and Ethics

US officials are also evaluating the national security implications for the emergence of DeepSeek, according to Karoline Leavitt, White House press secretary.

“I spoke to the National Security Council early today, and they are analyzing what these implications may be,” Leavitt said, reiterating statements by President Donald Trump that DeepSeek should serve as a warning to the US technology industry.

The announcement occurs after the US Navy prohibits the use of DeepSeek applications due to “potential security and ethics concerns.” According to CNBC, an email was sent to employees warning of the risks associated with the origin and use of the model.

Data security experts warn users to be careful with the tool as it collects large amounts of personal data and stores them on servers in China.

Meanwhile, DeepSeek claims to have been the target of cyber attacks. On Monday, the company announced that it would temporarily limit registrations due to “large-scale malicious attacks” on its software. A banner on the company’s website warns that registration may be congested because of the attacks.