Tom Gerken
Technological journalist
Getty images
Australia has prohibited Deepseek from all government devices and systems on what it says is the risk of security that the start-up of Chinese artificial intelligence (AI).
Deepseek amazed the world in January when he unveiled a chatbot that corresponded to the level of performance of the American rivals, while saying that he had a much lower training cost.
Billions of dollars have been destroyed on the stock market markets internationally, including in Australia, where the actions linked to AI – such as Tamias manufacturer, fell strongly overnight.
The Australian government has insisted that the ban is not due to the Chinese origins of the application, but due to the “unacceptable risk”, it poses to national security.
Deepseek was approached to comment.
Australia’s decision specifically obliges that government entities “prevent the use or installation of Deepseek web products, applications and web services”, as well as delete any previously installed system or system.
This means that a wide range of workers will not be able to use the country’s tools, including those working in various fields such as the Australia Electoral Commission and the Meteorology Office.
It is less clear if that means that Deepseek would be prohibited from computers in the public sector in different areas of the economy, such as schools.
The prohibition does not extend to the devices of individuals.
Growing – and familiar – concerns –
Western countries have a history of suspect in Chinese technology – notably the Huawei telecommunications company and the social media platform, Tiktok – which were both limited for national security reasons.
The initial reaction to Deepseek – which quickly became the most downloaded free application in the United Kingdom and the United States – seemed to be different.
President Donald Trump described it as “alarm clock” in the United States, but said that overall, it could be a positive development, if IA costs.
Since then, however, doubts about this have started to be expressed.
An Australian Minister of Sciences previously declared in January that the countries should be “very careful” on Deepseek, citing “data and confidentiality” concerns.
The application has already been prohibited in Italy, which temporarily blocked Chatgpt on confidentiality problems in March 2023.
The regulators in South Korea, Ireland and France have all started investigating how Deepseek manages user data, which he stores in servers in China.
White House press secretary Karoline Leavitt also said that the United States now examined the possible security implications.
The American navy would have forbidden its members to use Deepseek – although he did not confirm this at the BBC.
Generally, AI tools will analyze the prompts sent to them to improve their product.
This is true for applications such as Chatgpt and Google Gemini as much as Deepseek.
Everyone gathers and keep information, including email addresses and birth dates.
However, security experts have previously warned that anyone working on confidential security or national security areas must be aware of the risk that everything they enter in chatbots are preserved and analyzed by the developers of these tools.
