The Albanese government has been urged to reduce Australia’s “economic dependency” with China as the “good guys” and bad guys” of the artificial intelligence industry see Australia as “data harvesting haven”.

The AI industry has begun making waves in the Australian economy, with juggernaut Anthropic recently signing a massive $31.9 billion contract with the Queensland government.

However, the technology has become a complex new frontier for Australia’s national security, with growing concern that China’s repeated data breaches could become even more dangerous.

Strategic Analysis Australia Director Michael Shoebridge said Australia was “very badly equipped” to deal with the threat of AI, particularly from China.

“What worries me is that the government’s policy on AI looks a bit like how they’ve done the tax law changes … they’re proposing massive law changes because that’s what OpenAI and Anthropic want, but that will allow very high levels of data harvesting,” he told News24 host Caleb Bond.

“Australia will become a data harvesting haven, not just for the big four US AI companies but also for Chinese AI companies.

“So, we might think we’re helping the good guys, but we’ll also be helping the bad guys do mass data harvesting in Australia.”

However, US companies such as Google’s Gemini, OpenAI and Anthropic have been responsible for autonomously hacking security companies and government departments.

Mr Shoebridge said the Albanese government was maintaining Australia’s “economic dependency” on China, rather than the nation’s national security.

“Now, I get we trade with China, but how about trading more with some other countries as well so that China doesn’t get a bigger and bigger baseball bat on economics to us over the head,” he said.

Mr Shoebridge said as China deploys more cyber attacks in Australia, including breaches using artificial intelligence, Australia’s will be “silenced” as a hostage to economics.

Australia must focus on keeping local AI data as fears grow it could move offshore

Mr Shoebridge said AI-enabled hacks had occurred around the world just this week, with hackers heating hundreds of French supermarket fridges to as high as 50C, sparking fires and spoiling food.

“But another problem here is the big US companies, Anthropic and OpenAI, haven’t been able to control their own AI models. They’ve gone wrong,” he said.

“So you’ve got to be very careful about bringing them into top secret stuff here in Australia.

“And the problem I have with the way the government is managing China policy is they can never say anything to the Australian public … because they don’t want to offend Beijing.”

Shadow defence minister James Paterson underscored the threat of AI technology used by Chinese hackers and the importance of Australia developing suitable defence systems.

“Our intelligence agencies have publicly assessed that (China’s) cyber actors including the … Cyberspace Force and the Ministry of State Security’s cyber teams, have for years been identifying and seeking to exploit vulnerabilities in our critical networks,” Mr Paterson told the Centre for Independent Studies on Friday.

“The simple truth is that AI can identify and weaponise network vulnerabilities faster and more effectively than humans ever could.  Our only defence against AI-enhanced cyber-attacks is even better AI-enhanced cyber defence.”

Read related topics:ChinaAnthony Albanese