← Back to Stories

AI-driven surge in cyber threats and government response in Australia

By Updated 27 October 20262 articles from 2 independent sources

Consensus Summary

Australia’s cybersecurity landscape is under severe strain due to a surge in AI-powered cyber attacks, with government agencies warning of escalating threats to essential services. The Australian Signals Directorate (ASD) has documented a sharp rise in incidents, including a 20% increase in cybersecurity breaches in the 2025-26 financial year, alongside an 80% jump in potential malicious cyber incidents reported to external entities. The ASD’s annual threat report underscores how artificial intelligence is accelerating both the frequency and sophistication of attacks, forcing agencies to adapt rapidly. Among the most alarming developments was the first autonomous hack of a government website by a swarm of AI agents, which breached Australian health data security and prompted a rethink of collaboration with major AI companies like OpenAI and Anthropic.

The backdrop to this crisis is the rapid evolution of AI technology, which has given both criminal networks and state-aligned actors unprecedented tools to exploit vulnerabilities. The ASD’s report highlights how frontier AI models, such as Anthropic’s Mythos, were deemed too dangerous for public release due to their ability to identify unknown flaws in cyber defenses. This has forced the ASD to establish new protocols with AI labs to act as an early warning system, ensuring agencies like the ASD are engaged ahead of model testing. Meanwhile, the agency has also taken aggressive offensive measures, such as dismantling the Amadey malware network, which had infiltrated 180,000 computers globally and was sold on the dark web for as little as $US500. The ASD’s actions have prevented the theft of 5 terabytes of data, including 4.8 million files from Australians, demonstrating the scale of the threat.

Key figures in the response include ASD Director-General Abigail Bradshaw, who has emphasized the urgency of modernizing Australia’s cybersecurity infrastructure. Bradshaw warned that AI is not only a tool for attackers but also a critical asset for defenders, allowing the ASD to generate software capabilities more quickly. She stressed that legacy systems—outdated technology no longer supported by updates—remain the biggest risk, urging organizations to prioritize security over system availability, even if it means taking services offline temporarily. Defense Minister Richard Marles and Home Affairs Minister Tony Burke have echoed these concerns, framing AI as both a threat and an opportunity that Australia must harness to stay ahead of evolving cyber risks.

While both the ABC and the Guardian agree on the broad trends—such as the 20% increase in cyber incidents and the role of AI—they differ in specific details. The ABC focuses on the ASD’s offensive operations, including the disruption of the Amadey malware and the first autonomous hack of a government website, while the Guardian emphasizes the broader systemic risks, such as the 80% rise in potential malicious incidents and the need for collective action across industries. The Guardian also highlights the ASD’s warning about state-based actors in the Indo-Pacific preparing for potential conflicts, a framing absent from the ABC’s report. These differences reflect varying emphases: the ABC leans toward operational successes and immediate threats, while the Guardian underscores long-term strategic vulnerabilities and the need for systemic change.

Looking ahead, the ASD and government officials have made it clear that the situation demands urgent action. Bradshaw has called for accelerated modernization of technology and cybersecurity foundations, while also stressing the importance of collaboration with AI companies and international partners. The recent vulnerabilities in PHP, which triggered a rush to patch systems, serve as a reminder of how quickly threats can escalate. With AI continuing to reshape the threat landscape, the ASD’s message is clear: Australia must act now to secure its digital future, or risk facing even greater disruptions in the years to come.

✓ Verified by 2+ sources

Key details reported by multiple sources:

  • The Australian Signals Directorate (ASD) reported a 20% increase in cyber security incidents in the 2025-26 year compared to the previous year.
  • The ASD responded to 4,400 potential malicious cyber incidents in the 2025-26 year, an 80% increase from the previous year.
  • The first autonomous hack of a government website occurred in the past year, involving a swarm of AI agents breaching Australian health data security.
  • The ASD disrupted a criminal network using the Amadey malware, which had access to 180,000 computers globally, including in Australia.
  • The ASD destroyed 5 terabytes of stolen data, including 4.8 million files from Australians and hundreds of thousands of files from other victims.
  • The ASD warned of elevated threat levels due to AI-powered cyber attacks, emphasizing the need for modernizing cybersecurity infrastructure.
  • The ASD has established protocols with AI companies, including OpenAI, to act as an early warning system for emerging threats.
  • The ASD’s annual cyber threat report highlights AI as a major factor in increasing the scope and scale of cyber attacks.

Points of Difference

Details reported by only one source:

ABC News
  • The ASD revealed that the release of frontier AI models like Anthropic's Mythos redefined cyber threats, as it was deemed too dangerous to release publicly due to its ability to find vulnerabilities in cyber defense systems.
  • The ASD took down criminal groups selling Amadey malware, which was found hidden in a free, illegitimate version of Photoshop editing software and sold on the dark web for about $US500.
  • The ASD reported a 20% jump in cyber security incidents in the last financial year, with the first autonomous hack of a government website occurring in the past year.
  • The ASD worked 'very quickly' to set up protocols with AI companies, including OpenAI, to act as an early warning system.
The Guardian
  • The ASD responded to 1,500 cybersecurity incidents in the 2025-26 year, a 20% increase from the previous year.
  • The ASD notified external entities of over 4,400 potential malicious cyber incidents, up 80% in a year.
  • The ASD warned that state-based actors in the Indo-Pacific could be launching attacks to prepare for potential future conflicts, including intelligence collection or disruption activities.
  • The ASD urged organizations to prioritize system security over availability, acknowledging that more services may need to go offline for security updates.

Where the reporting differs

Details that conflict, or appear in only some outlets:

  • The ABC states the ASD reported a 20% jump in cyber security incidents 'last financial year,' while the Guardian specifies the increase occurred in the '2025-26 year,' which may imply different reporting periods.
  • The ABC mentions a 20% jump in incidents 'last financial year' with no specific year, while the Guardian provides a concrete figure of 1,500 incidents in the 2025-26 year, which is not directly comparable.

Source Articles

ABC

Top cyber spy braces for more AI-powered hacking attacks

Australia's cyber-spy agency offers an insight into how it is dealing with the surge in AI threats.

GUARDIAN

Australians should prepare for online services to go down unexpectedly as threats rise, cyber agency warns

Australian Signals Directorate chief says advanced AI tools are needed to defend against ‘AI-enabled attackers’ Get our daily political email , free app or daily news podcast Australians should prepare for more basic and important online services to go offline unexpectedly, the nation’s signals intelligence agency has warned, urging that “system security must be prioritised over system availability”. The Australian Signals Directorate (ASD) also says businesses and critical infrastructure must u

More Technology stories

Latest cross-verified stories

Browse all stories from October 2026 in the archive.