Big tech’s national security claim challenged in Australia
A defence report commissioned by the Australian Recording Industry Association says technology companies are presenting commercial demands as national security…
Former intelligence official Michael Shoebridge says big tech’s national security case for weakening Australian copyright laws is a commercial argument, not a proven security need. A report commissioned by the Australian Recording Industry Association warns that loosening protections could also make Australia more attractive to cybercriminals.
The debate puts pressure on Prime Minister Anthony Albanese’s government as OpenAI and Anthropic seek greater access to copyrighted works for AI training. The companies have linked local investment and training plans to changes in the law, while creative industries oppose weakening existing protections.
That could affect what creators are paid for their work and how AI developers gain access to Australian books, music and art. It also raises a security question: whether training models locally offers Australia control of the technology or simply gives developers a more permissive legal environment.
What does the report say about big tech’s national security argument?
The report argues that claims about national security do not establish a need to train AI models on copyrighted creative works. Shoebridge, a former senior intelligence and defence official, says cybersecurity relies on specialised code, encryption and classified intelligence instead.
“AI for national security does not need Powderfinger’s back catalogue,” Shoebridge said. He added that companies should show how training on Australian songs, novels or paintings would help prevent a cyberattack on the power grid or disrupt a terrorist plot.
His report says defence and intelligence agencies can access advanced AI through commercial arrangements. Training American-owned models in Australian data centres, it argues, would not give Australia ownership or control of the technology.
There is a different view inside government. At a parliamentary inquiry in September, Australian Signals Directorate director-general Abigail Bradshaw said local training of advanced AI models would bring “significant benefits for national security.” She argued that countries building capability early would be better placed to shape future technologies and international standards.
The Attorney-General’s Department told the same inquiry that copyright was the main concern raised by major AI developers. OpenAI vice-president Ann O’Leary said the company could not establish an Australian training centre under existing laws. Anthropic executive Jeff Bleich described individual content licensing as “technically impossible”.
Why are copyright rules linked to security concerns?
Shoebridge warns that relaxing copyright protections could draw Chinese AI developers, organised crime networks and hackers seeking to exploit more permissive laws. The report contrasts that risk with what it describes as stronger protections and creator-payment requirements in other advanced economies.
It also questions the security record of major AI developers. OpenAI acknowledged that its autonomous agents had accessed Australian government websites without authorisation, including systems involving Medicare-related information.
The report’s argument is contested by officials who see local AI capability as a security priority. But the disagreement reaches beyond where models are trained: it concerns who controls the systems, what material companies can use, and whether creators receive payment.
Australia’s creative sector, including musicians, authors and media companies, has united against a move to loosen existing laws. Shoebridge said: “If the companies believe it does, the burden is on them to prove it.”
