Four in five Australian organisations now let artificial intelligence (AI) make decisions and take actions that required human approval a year ago, but only 55% can see and audit every action AI agents take, according to new research from cybersecurity firm Illumio.
The findings come from a global survey of 1,100 senior IT security decision-makers, including respondents in Australia. Illumio’s worldwide results put the equivalent figures at 70% and 49%.
Australia recorded the highest use of AI in cybersecurity operations among the markets surveyed. The research found 80% of Australian businesses use AI in cybersecurity, compared with a global average of 67%.
That use is extending to autonomous tasks. One-third of Australian organisations said they allow AI to grant or revoke access permissions without human review, while 28% let it run security investigations and 26% permit unsupervised code changes.
Nearly three-quarters of Australian businesses, or 73%, said they had experienced at least one AI-related security issue or warning sign in the past 12 months. Among the issues reported, 35% said AI had accessed data it should not have, 33% said AI made a recommendation that would have created a security risk if followed, 30% said AI tools or agents connected to systems without security approval and 28% said AI produced or triggered an unauthorised action.
Christopher Horlyck, head of cyber security resilience at the Australian Signals Directorate (ASD), said the findings underline the need for tighter controls.
“As AI is used more extensively in cybersecurity operations, Australian organisations must ensure this does not come at the expense of security and control. AI agents with excessive access or insufficient controls may take unintended or unauthorised actions,” Horlyck said.
He pointed to ASD guidance on agentic AI harnesses, published Sept. 11, which advises restricting access, maintaining monitoring and audit logs and applying human oversight to high-impact actions.
The research also suggests some organisations are more willing to accept cyber risk for the perceived benefits of AI. Globally, 53% of respondents said they would accept more cyber risk than 12 months ago, SecurityBrief Australia reported. Within that group, more than half allow autonomous AI actions in non-production environments, and a further 45% allow them in production, where systems are live.
Only 37% said frontier AI had made network segmentation or microsegmentation more important, and 26% named explicitly defined permitted actions as one of their two main controls. Illumio said the mismatch leaves organisations exposed because AI systems can act with legitimate access rather than breaking in from outside.
“Australian organisations are putting AI in the security driver’s seat, but many can’t see or audit every action it takes. That’s a dangerous gap. Unlike external hackers, AI agents can operate with legitimate access to sensitive systems and data, making their actions harder to distinguish from normal activity,” said Andrew Kay, senior director of systems engineering, APJ, at Illumio.
“Security teams cannot predict every action an agent might take, so they must limit what it can reach and be able to contain the damage when it behaves unexpectedly.”
The research follows an April 30 letter from the Australian Prudential Regulation Authority (APRA) to all regulated entities, based on a review of large banks, insurers and superannuation trustees, Insurance Business reported. APRA found identity and access management systems had not been updated for non-human actors such as AI agents. APRA said it would pursue enforcement where entities fail to manage AI risks proportionate to their size and complexity.
Separate Semperis research, released mid-May, found 21% of Australian organisations were certain they could recover control of their systems if AI exposed administrator credentials, compared with 32% globally, the outlet reported.
