Chinese AI chatbot DeepSeek to be banned from Government devices over national security concerns

Emerging Chinese-owned and operated AI tool DeepSeek will be banned from all Australian Government devices immediately over national security concerns.

National security and intelligence agencies have determined the chatbot platform, developed by a Hangzhou-based startup, “poses an unacceptable risk to Australian Government technology”.

The direction was made under the same framework used to prohibit TikTok from Government devices in 2023.

Home Affairs Minister Tony Burke said the Government was “taking swift and decisive action to protect Australia’s national security and national interest”.

“AI is a technology full of potential and opportunity – but the Government will not hesitate to act when our agencies identify a national security risk,” he said.

“Our approach is country-agnostic and focused on the risk to the Australian Government and our assets.”

The Government has also encouraged “all Australians” to review how the platform is using their data.

The announcement comes after Australia’s largest cybersecurity provider CyberCX issued a threat advisory on Friday, recommending the platform be banned from Government devices because it is “almost certain” that DeepSeek and the user data it collects are subject to direction and control by the Chinese government.

Taiwan banned Government agencies from using the app this week, as did a NSW Government department, while Italy’s data protection authority blocked its use to protect Italians’ data while it investigated the companies behind DeepSeek.

According to DeepSeek’s own privacy policy, it collects large amounts of personal data — including a user’s date of birth and phone number, and ‘keystroke patterns’ — from users, stored “in secure servers” in China.

Dana Mckay from RMIT said the concern was that the data Chinese apps like DeepSeek collected was made available to the Chinese Government for a miriad of reasons.

“It is fair to ask whether DeepSeek is more dangerous to Australian national security than, say, OpenAI which collects similar data: the difference is that OpenAI will only give data to government to comply with relevant laws, and this typically means where a crime may have been committed,” she said.

DeepSeek triggered a global frenzy last week, after it claimed its models functioned on par with its US competitors like OpenAI (which makes ChatGPT) and advanced Meta models, at a fraction of the cost.

US chip giant Nvidia’s share price dropped by roughly 17 per cent last Monday, wiping almost $1 trillion off its market value.

The US Commerce Department is investigating whether DeepSeek has been using US chips that are not allowed to be shipped to China, weekend reports suggested

DeepSeek has said it has used computer chips from Nvidia but the department is believed to be investigating whether restrictions on Nvidia’s most sophisticated AI chips from reaching China have been flouted.