One of Australia’s leading intelligence agencies has issued a warning that artificial intelligence (AI) attacks could exploit the nation's older technology. This comes as the Australian government is in discussions regarding regulations for the rapid development of AI.
Abigail Bradshaw, the director general of the Australian Signals Directorate, stated at the Sydney Dialogue summit in Canberra that her agency does not know the exact number of active AI agents on the internet, but believes there are "a lot."
Bradshaw emphasised that Australian governments and businesses face the prospect of spending "enormous" sums to update their vulnerable, outdated technology. This would involve taking offline systems that Australians currently expect to be constantly available.
The Greens party called for mandatory safety testing and transparency obligations for advanced AI systems, but their motion was rejected by Labor senators. Andrew Charlton, the assistant technology minister, acknowledged the warnings from companies but indicated that the Albanese government would set its own forthcoming guidelines for AI companies, which are expected early next year.
Anthropic's chief lawyer, Jeff Bleich, expressed that the company has been "heartened" by support for its calls for regulation and believes Australia should play a role in setting these rules. However, Bleich also stated that Anthropic believes rules are needed within months due to the technology's exponential development.