What countries have prohibited Deepseek and why? | Commercial and economic news

This week, government agencies in countries such as South Korea and Australia have blocked access to the new Chinese artificial intelligence chatbot program (AI), the new Deepseek IA Chatbot program, mainly for government employees.

Other countries, including the United States, said they could also try to block Deepseek mobile devices from government employees, media said. All cite “security problems” concerning Chinese technology and a lack of clarity on how the personal information of users is managed by the operator.

Last month, Deepseek made the headlines after provoking the drop in stock prices for American companies, after having claimed that his model would only cost the money fraction that his competitors had spent on their own programs in programs ‘IA to build. The news joined the users of social media: “I cannot believe that Chatgpt has lost its job at AI.”

Here is what we know about Deepseek and why countries prohibit it.

What is Deepseek?

Deepseek is a startup of Chinese AI. Based in Hangzhou City, it was founded by the entrepreneur and businessman Liang Feng in 2023. Liang also founded the group of roofing funds of $ 7 billion with two university classmates in 2016.

In January, Deepseek published the latest model of its program, Deepseek R1, which is a free chatbot fueled by AI with a very similar look and feeling to Chatgpt, belonging to Openai, whose seat is in California.

AI chatbots are computer programs that simulate human style conversation with a user. Users can ask questions in the bot and then generate conversational answers using information to which he has access to the Internet and with which he was “trained”.

Among a plethora of potential uses, these programs can be used to solve mathematics problems, a text project such as emails and documents, and translate or write codes.

How does Deepseek differ from Chatgpt and other similar programs?

On the one hand, its developers say that it is a lot, much cheaper to build.

Deepseek has sent shock waves in AI circles when the company published an article in December declaring that “training” of the last depth model – organization and enracing the information it needs to answer questions – would require Less than $ 6 million in NVIDIA H800 chips calculation power.

This is only a small fraction of IA budgets of several billion dollars enjoyed by American technology giants such as Openai for Chatgpt and Google belonging to the United States for Gemini. OPENAI CEO Sam Altman said that the latest OpenAI program GPT-4 cost more than $ 100 million to train, Wired reported in April 2023.

The new Deepseek program was published to the public on January 20. On January 27, the Deepseek application had already reached the top of Apple App Store. Consequently, the American multinational NVIDIA, which maintains a quasi-monopoly on the manufacture of semiconductors for a generative AI, lost nearly $ 600 billion in market capitalization when the share price dropped by 17%.

What countries prohibit the Deepseek AI program?

Some government agencies in several countries are looking for or promulgated for AI software for their employees. These include:

UNITED STATES

The Wall Street Journal reported on Thursday that US legislators planned to introduce a government bill to block Deepseek of the government belonging to the government.

On January 31, the American space agency NASA blocked Deepseek from its systems and the devices of its employees. A week earlier, the American navy warned its members in an email against the use of Deepseek due to “potential security and ethical concerns associated with the origin and use of the model”, reported CNBC.

South Korea

A spokesman for the Ministry of Trade, Industry and Energy of South Korea announced on Wednesday that the Ministry of Industry had temporarily prohibited Deepseek on employee devices, also citing security problems.

Officials said the government had exhorted ministries and agencies on Tuesday paying attention to the use of AI programs in general, including Chatgpt and Deepseek.

This intervened after the Déagon Information Private Information Commission for the Personal Information Commission, announced on January 31 that it would send a written request to Deepseek for more details on the Management of personal user information.

Korea Hydro & Nuclear Power, which is led by the South Korean government, said that it had blocked the use of AI services on its workers’ devices including Deepseek last month.

Australia

The Australian government announced on Tuesday that it had blocked access to Deepseek on all government devices, saying that there were “security risks”.

This ban was mandated for all government agencies in a Tuesday declaration by the secretary of the Department of Internal Affairs. The declaration has ordered all government entities to “prevent the use or installation of Deepseek web products, applications and services and, when noted, delete all existing bodies of products, applications and services Deepseek of all Australian government systems and devices ”.

The Minister of Internal Affairs, Tony Burke, said that the ban was “to protect national security and the national interest of Australia,” the Australian media reported.

Italy

On January 30, the Italian Data Protection Authority (Guarantor) announced that it had ordered “the limitation of processing of Italian users” by Deepseek because of the lack of information on how Deepseek could use the personal data provided by users.

Two days earlier, the guarantor had announced that it was looking for answers on how user data was stored and managed by the Chinese startup.

Taiwan

Taiwan prevented government services from using Deepseek programs on Monday, also blaming security risks.

Why did some countries have placed prohibitions on the use of Deepseek?

Most countries blocking Deepseek programs say they are concerned about the security risks made by Chinese demand. They also say that they do not have enough information on how the personal data of users will be stored or used by the group.

According to Deepseek Privacy PolicyIt collects the following user data:

• Personal information, including email, telephone number, password and date of birth, which is used to register for the request.
• Cat history in the application, including text or audio that the user enters the chatbot.
• Technical information on the user’s device and network, such as the IP address, the typing models and the operating system.

He shares this information with service providers and advertising partners. This information is kept during “as long as necessary,” said the company’s website.

According to Chatgpt Privacy PolicyOPENAI also collects personal information such as the name and contact information given during registration, device information such as IP address and chatbot entries “only as long as we need it”. This information can also be shared with Openai affiliates.

On Wednesday, ABC News quoted a report by Ivan Tsarynny, CEO of Feroot Security, a cybersecurity company based in Ontario which affirmed that Deepseek “A of hidden code in its programming which has the integrated capacity to send user data directly to the Chinese government ”.

Tsarynny told ABC that the Deepseek application is capable of sending user data to “CMPASSPORT.com, the online register of China Mobile, a telecommunications company held and operated by the Chinese government”. Al Jazeera has not been able to independently check this assertion.

Is it an unusual practice?

“Practically all major technological companies – from Meta to Google in Openai – operate user data to a certain extent,” Al Jazeera Eddy Borges -Rey, an associate professor in the Northwestern University in Qatar, told Al Jazeera.

“They use data for targeted advertising, algorithmic refinement and AI training. Many have been sentenced to a fine or an investigation into privacy violations, but they continue to operate because their activities are somewhat regulated in courts like the EU and the United States, “he added .

Borges-Rey explained that Chinese platforms such as Deepseek are “treated differently by the West because they are considered to be operating under the jurisdiction of the Chinese government, which has laws (such as national intelligence law) which theoretically allow State access to corporate data ”.

He added: “Western governments fear that user data collected by Chinese platforms can be used for espionage, influence of operations or surveillance. That this happens in practice is questionable, but the simple possibility is sufficient to justify the prohibitions from the point of view of national security. »»

On the other hand, Western requests are not perceived as a threat of national security by Western governments. “Western companies are often considered problematic but repairable thanks to regulations, while Chinese companies are considered to be direct security threats requiring prohibitions.”

In 2023, Chatgpt expressed her concerns that she had raped the General Data Protection Regulations of the European Union (GDPR). On April 1, Italy temporarily blocked the service for all users in the country. On April 28, 2023, Chatgpt was restored in Italy and Openai said that he had “tackled or clarified” the problems raised by the Guarantor. The Guarantor told the BBC that it “hosted open measures”.

For its part, China blocks access to Western technological companies, especially X, Facebook and even Chatgpt for all users in the country.