Australia bans Chinese AI firm DeepSeek from government systems over security risks

Australia's federal government has moved to strip DeepSeek from every government computer and device, citing national security risks tied to the Chinese artificial intelligence company. The ban applies to all products, applications, and services offered by the Beijing-based firm, effective immediately across federal agencies.

The decision comes from Home Affairs Minister Tony Burke, who framed it as a straightforward security measure rather than a broader statement about Chinese technology. "AI is a technology full of potential and opportunity," Burke said, "but the government will not hesitate to act when our agencies identify a national security risk." He emphasized the ban was "country-agnostic and focused on the risk to the Australian government and our assets"—language designed to position the move as pragmatic rather than ideological.

DeepSeek arrived in the global consciousness in January 2025 as a genuine shock to the technology industry. The Chinese startup built an AI chatbot that produces responses comparable to established competitors like OpenAI's ChatGPT, but at a fraction of the cost and using significantly less computing power. The implications rippled through financial markets: investors immediately questioned whether other AI companies had been overvalued, triggering a sell-off that erased nearly $600 billion from Nvidia's market capitalization in a single trading day. That kind of market disruption tends to draw government attention, particularly when the disruptor is based in China.

Australia is not alone in its caution. Taiwan's government banned its own departments from using DeepSeek services over information security concerns. Australia itself has precedent for this kind of action: in 2023, it prohibited TikTok from government-issued devices, citing similar risks to sensitive data. The pattern suggests a hardening posture toward technology platforms based in or controlled by China, at least in contexts where government data is at stake.

What makes the DeepSeek ban notable is what it does not do. The Australian government has not imposed a blanket prohibition on American AI services, despite documented risks. ChatGPT, made by OpenAI, has been flagged for potentially incorporating sensitive information from user prompts into its training dataset—information that could later be exposed. The Department of Home Affairs suspended access to both ChatGPT and Google's AI assistant from its systems in May 2023, but no government-wide ban followed. The then-secretary of the Department of Home Affairs, Mike Pezullo, called for a whole-of-government approach to the question of whether to deploy these technologies at all, but no such unified policy materialized.

For ordinary Australians, the government's position is different. Employment Minister Murray Watt said individuals remain free to download and use DeepSeek if they choose. "As a government, we're leaving the choice up to individuals for what apps they have on their phone," Watt told Seven on Wednesday. "But the government has decided it is not secure to have on government phones and people might want to have a think about that." The government plans to issue further guidance to help citizens understand the privacy implications of using the service, but no legal restriction on public use is being contemplated.

The ban sits within a broader, still-forming landscape of AI governance in Australia. In September 2024, the federal government released voluntary guidelines on AI use across the country, signaling an intent to shape the technology's deployment without heavy-handed regulation. The DeepSeek decision suggests that intent has limits—when national security is the concern, and when the company is Chinese, the government will act decisively. What remains unclear is whether this will be the first of a series of restrictions, or whether it reflects a specific assessment of DeepSeek's particular risks.