The burst of the R1 conversational robot of the Chinese start-up Deepseek amazed the tech industry by its ability to match its American competitors at a lower cost, rebutting the cards of artificial intelligence (AI). But this barely launched chatbot, several governments have already restricted its use, invoking threats to national security or potential leaks of sensitive information.
Who forbidden Deepseek?
The Talie is the first country to have opened an investigation into Deepseek, to whom it prohibited to process Italian user data. In 2023, the Italian surveillance organization had already temporarily blocked the Openai chatbot in the country, Chatgpt, for reasons of confidentiality.
After Italy, the Taiwanese authorities prohibited their civil servants and key infrastructure to use the applications of the Chinese start-up, invoking risks for “national information security”. This decision was followed a few days later by Australia. “After an analysis of risks and threats, I considered that the use of product products, applications and internet services of Deepseek represented an unacceptable level of security for the Australian government,” wrote the Minister of The interior, Stephanie Foster, in a directive published on February 4.
In South Korea, several ministries, including the one who oversees relations with North Korea, blocked access to Deepseek on their computers. These blocking measures also concern “military PCs”. This Monday, February 17, the South Korean authorities also announced the withdrawal of Deepseek from local application stores while studying the way in which the Chinese startup manages user data. Deepseek “recognized certain shortcomings” concerning compliance with privacy, said the South Korean commission responsible for the protection of personal data.
In the United States, a bill was presented by parliamentarians to prevent the use of Deepseek – qualified as a “affiliated company to the Chinese Communist Party” by representative Darin Lahood – on government devices for cybersecurity reasons .
What are these countries fear?
The general conditions of Deepseek contain a section on the transmission of personal data to third parties, very similar to that of Chatgpt, the conversational robot of its American rival Openai. But “In China, when the government requests access to user data, companies are legally required to provide them,” notes Youm Heung-Youl, teacher and data security specialist at Soonchunhyang University in South Korea .
According to Deepseek’s privacy policy, the Chinese start-up also collects information on “keyboard strikes”, that is to say any interaction that a user performs with a touch of his keyboard.
The CNIL in France, responsible for data protection, and the Irish Commission (DPC) Data Commission (DPC) – which regulates technological giants on behalf of the EU- have claimed explanations from Deepseek on data processing.
Are these concerns justified?
“No major tech company is politically neutral,” said Vladimir Tikhonov, professor of Korean studies at the University of Oslo. “Google stores data relating to navigation history and it is naive to think that they are not shared with government agencies when they request it”. Vladimir Tikhonov recognizes, however, that the collaboration of large tech companies with the Chinese authorities is probably “more in -depth”.
Beijing claims that the Chinese government “will never require businesses or individuals that they collect or illegally store data”. China has denounced the restrictions recently imposed by several countries, seeing it as a “politicization of economic, commercial and technological issues”.