Overview
- The Australian Signals Directorate published its Annual Cyber Threat Report on Tuesday, October 27, saying AI has reshaped the threat and the agency handled more than 82,000 cybercrime reports, responded to over 1,500 incidents, and issued about 4,400 notifications in 2025–26.
- The report says frontier and agentic AI now finds and weaponises software flaws far faster than before, helping attackers discover critical PHP vulnerabilities that forced mass patching and contributing to autonomous agent breaches of government and health data.
- ASD says state‑aligned actors linked to China, Russia and North Korea routinely target Australian organisations to steal information and preposition for disruption, and the agency used offensive operations to sever an Amadey malware controller and destroy roughly 4.8 million stolen Australian files.
- The agency is urging organisations to prioritise security over availability by taking systems offline to patch, remove exposed credentials, rotate secrets, isolate test agents from production, extend logging and work directly with major model operators and Five Eyes partners.
- Officials warn the human costs could be large: regulators note AI‑amplified attacks raise risks to everyday services and to Australians’ savings held in $4.5 trillion of superannuation, while businesses face acute cyber skills shortages that leave many systems vulnerable.