EXCLUSIVE: A New South Wales government department has blocked Chinese artificial intelligence chatbot DeepSeek from its devices and systems over privacy concerns, as other states begin to investigate and monitor the generative AI system.

DeepSeek, which caused havoc with American technology stocks as its use skyrocketed last month, was purportedly created at a much lower cost and with less computing power than US contemporaries, such as OpenAI’s popular ChatGPT.

The Large Language Model (LLM) has attracted concern from some Western nations — including Australia — because the data it collects is stored in China, where companies must comply with data requests from the Chinese government.

While state and territory government employees across the country have previously been told not to enter sensitive or personal data into AI chatbots, some departments are taking extra precautionary measures around DeepSeek.

The New South Wales Department of Customer Service told staff on Thursday 30 January that it had made the decision to block access to DeepSeek on its devices and systems.

"After careful analysis of the way the platform handles data and accesses technical information, we have made the decision to block access to this service on work devices and systems to protect our data and information,” read a note to staff.

The department told employees it was working with other NSW government departments on the issue.

In a statement to Information Age, the department confirmed it had "taken a precautionary approach to restrict corporate access to DeepSeek AI".

In Victoria, some staff at the state’s Department of Government Services have been informally told they should not use DeepSeek on work devices, Information Age understands.

In a statement, the department said it was “currently looking into DeepSeek” but added Victorian Public Sector staff had already been told not to input sensitive information into generative AI systems.

The office of the South Australian Chief Information Officer said it was working with the federal government and state and territory governments on monitoring systems such as DeepSeek, but did not confirm any ban in that state.

The Tasmanian government said it had not imposed any whole-of-government restriction on DeepSeek but was "committed to the safe and responsible use of AI ".

An ACT government spokesperson said the territory’s Cyber Security Centre “continuously monitors emerging threats and risks associated with software platforms and online services”.

Queensland's Department of Customer Services, Open Data, and Small and Family Business, Western Australia’s Office of Digital Government, and the Northern Territory’s Department of Corporate and Digital Development did not respond by deadline.

Governments warned about DeepSeek’s China links

Prominent Australian cybersecurity company CyberCX published a warning about DeepSeek on Friday, stating, “We assess it is almost certain that DeepSeek, the models and apps it creates, and the user data it collects, is subject to direction and control by the Chinese government.”

The company recommended all organisations — but "especially critical infrastructure organisations, government departments and agencies and organisations storing or processing commercially sensitive or personal information" — strongly consider restricting access to DeepSeek.

CyberCX said it had “high confidence” DeepSeek’s AI model complied with Chinese government censorship requirements and stored information collected from users on servers in mainland China.


DeepSeek's AI model is available as an app, as well as open-source software which can be run locally on consumer devices. Photo: Shutterstock

Last Tuesday, Minister for Industry and Science, Ed Husic, argued many Australians would be suspicious of DeepSeek’s data and privacy protocols, but said it was “too early to jump to conclusions” such as banning the app in Australia.

On Wednesday, Treasurer Jim Chalmers added the government “would urge Australians to be cautious about this new [DeepSeek] technology”.

The Department of Home Affairs, which helps develop cybersecurity policy for the Australian government, did not respond to a request for comment.

The federal government banned Chinese-owned social media app TikTok from being used on government devices back in April 2023.

Global powers react to privacy and security concerns

Several US government departments and agencies have moved to block access to DeepSeek.

The US Department of Defense shut off access to DeepSeek last week, reportedly at least two days after several employees connected their work computers to the company's servers in China, according to Bloomberg.

The US Navy also issued a warning to its members to avoid using DeepSeek, due to “potential security and ethical concerns”.

Elsewhere, democratically governed Taiwan banned government departments from using DeepSeek on Monday, citing security risks.


DeepSeek rose to the top of global app store charts in January, overtaking competitors such as Google's Gemini and OpenAI's ChatGPT. Photo: Shutterstock

In Europe, Italy’s data protection authority blocked access to DeepSeek on Thursday and said it was not satisfied with responses it received from the company about the protection of personal data.

Privacy watchdogs in the likes of South Korea, France, Ireland, and the Netherlands said they also planned to investigate DeepSeek and ask the company how the personal information of users was managed.

Australian watchdog monitors 'with deep interest'

The Office of the Australian Information Commissioner (OAIC) said while it did not have plans to launch its own investigation into DeepSeek, it watched its international counterparts “with deep interest” and was "in regular communication with our peers on this and other issues”.

“Given the complexity of the AI sector and the resource imbalance between AI companies and privacy regulators, we are looking to avoid duplication, rather than replicate it, and instead we are seeking to use our powers to cover the field,” the OAIC told Information Age.

“Our job is to use the right tool in our toolbox in a proportionate and responsive way to ensure the best privacy outcomes for Australia.”

The OAIC flagged it “may” still launch investigations and enforcement actions in the future.