‘Same as TikTok’: Cybersecurity warning for AI app DeepSeek

Australian cybersecurity executives and the federal opposition have sounded a warning about Chinese artificial intelligence start-up DeepSeek, which has rocketed to the top of app stores globally and roiled sharemarkets with its threat to America’s AI dominance.

DeepSeek’s rapid ascent this week took investors and analysts by surprise – as well as apparently the company itself, which said it would temporarily limit user registrations due to repeated outages as it struggled to keep up with its sudden popularity.

DeepSeek’s rapid ascent this week took investors and rivals by surprise, as well as seemingly the company itself.Credit: Bloomberg

Its rampant rise also triggered a bloodbath on Wall Street, dragging down shares in American AI giant Nvidia in the biggest fall in US stock market history, as well as those of Oracle, Meta and Google parent company Alphabet.

Little is known about DeepSeek, a small Hangzhou-based start-up founded in 2023 by entrepreneur Liang Wenfeng, who runs a hedge fund, High-Flyer Capital, which uses AI to identify patterns in stock prices.

Last week, DeepSeek launched its free AI chatbot, which its creators say beats the likes of ChatGPT in terms of technical capabilities with a fraction of the energy usage and cost. It’s now dominating app store charts globally.

Australian cybersecurity leaders and the federal opposition are alarmed, however, that the company is censoring content and sending sensitive data to Chinese servers, echoing similar concerns about TikTok and network infrastructure provider Huawei.

Liberal senator James Paterson has previously urged Australia to ban TikTok, and said Australians should be cautious when using Chinese-owned apps such as DeepSeek.

“It is clear from the app’s terms and conditions that user data is collected in abundance, including what prompts are entered into the app,” he said.

“The app’s policies also make it clear they will hand over this information if requested to do so by the Chinese government. I am also concerned about the bias the AI algorithm appears to have towards a Chinese Communist Party worldview, as it has been shown refusing to provide information on topics sensitive to the CCP and its historical oppression of minorities and democratic activists.

“We must make sure this does not become another tool for authoritarian propaganda, and I hope the Albanese government is closely co-ordinating a response with our closest allies.”

Industry and Science Minister Ed Husic refused to indicate whether DeepSeek was a national concern when asked several times at a press conference, saying the government would “keep an open mind” and watch for anything that presented a risk to the national interest.

“I think it’s too early to jump to conclusions,” he said. “We will clearly be informed by the advice of National Intelligence Community in relation to threats as they might present.”

He told ABC Radio: “We’re going to see more and more of these products being made.

“People will put these new products through their paces, but there has been a long march towards the greater use of AI [...] and I just believe that’s going to continue well into the future. If anything, it will accelerate.”

Industry and Science Minister Ed Husic says the production of artificial intelligence systems such as DeepSeek will continue to be ramped up.Credit: Alex Ellinghausen

Alastair MacGibbon, Australia’s cybersecurity tsar under former prime minister Malcolm Turnbull, said Australian users should be concerned about censorship and handing over sensitive data to the Chinese state.

The cybersecurity veteran, now chief strategy officer at cybersecurity firm CyberCX, said users needed to ask themselves whom they trusted with their information.

“Anybody using DeepSeek or other Chinese large language models should try asking it about Tiananmen Square, pro-democracy movements in Hong Kong or for criticisms of Xi Jinping. If it can’t or won’t answer these questions, ask yourself what else will it not tell me?,” MacGibbon said.

“While for the most part, Western tech companies have learned to be cautious about who can access IP and user information, there are no similar conversations happening in China … Anything we enter into DeepSeek is going straight into the hands of the Chinese Communist Party.

“The conversations we have had about Huawei and TikTok are the same conversations we will have about DeepSeek. Internet users and governments need to be alert to this.”

DeepSeek’s rise has also sparked questions about the billions of dollars already invested by US giants on AI innovations. The Chinese start-up said it spent just $US5.6 million ($8.9 million) on computing power for its base R1 model, relying on less powerful H800 chips produced by Nvidia, which have been banned from export to China since October.

In an interview with Chinese media, DeepSeek’s founder Liang said that “AI should be affordable and accessible to everyone”.

Leading US venture capitalist Marc Andreessen compared the launch of DeepSeek’s R1 model to a pivotal moment in the US-USSR Cold War space race, describing it as AI’s “Sputnik moment”, while OpenAI chief executive Sam Altman said R1 was impressive “particularly around what they’re able to deliver for the price”.

Chief scientist Tony Hayment said Australia’s record of governance meant the country would probably be in good shape to combat some of the dangers of AI, but DeepSeek showed how quickly it could develop: “DeepSeek is genuine, and is really going to change our lives, at least change the stock market.

“But it shows you the pace at which innovation and science and technology goes privately funded in Shanghai, by a bunch of talented 22-year-olds, without access to the world’s best chips, without access to Nvidia chips, seem to have created something that’s even better that the best companies in the Western world have done.”

DeepSeek did not return a request for comment.

Maria Zhang, an associate professor at the University of Sydney, said DeepSeek was part of a new wave of Chinese start-ups defined by an ability to maximise limited resources.

Until now, it had been presumed that billion-dollar investments and access to the latest generation of specialised Nvidia processors were prerequisites for developing state-of-the-art AI systems.

“The West has missed the emergence of China’s Innovation 2.0, characterised by a new generation of entrepreneurs who prioritise foundational research and long-term technological advancement over quick profits,” she said.

The Market Recap newsletter is a wrap of the day’s trading. Get it each weekday afternoon