Guest Articles / Guest Opinion
How Government Agencies Can Respond to Faster AI-driven Cyber Attacks
GUEST OPINION: In April, Anthropic revealed that its Mythos preview model identified thousands of zero-day vulnerabilities in major operating systems and web browsers, including a 27-year-old bug in OpenBSD. This has sparked intense discussion across the cybersecurity community.
For Australian CISOs at the federal, state, and local government levels, the implications are significant. Government agencies face a different threat profile from most private sector organisations. They are high-value targets for nation-state actors and other sophisticated adversaries that have the resources to develop or acquire advanced AI-assisted attack tools.
These adversaries are well-funded, highly skilled, and focused on critical systems and sensitive information. The Australian Signals Directorate’s Australian Cyber Security Centre (ACSC) has repeatedly warned that state-backed actors continue to target Australian government networks and critical infrastructure. With AI, attackers are finding vulnerabilities faster, and the time between discovery and exploitation is shrinking. This places greater pressure on agencies to identify and address risks before they can be exploited.
The gap between AI-enabled attackers and public sector defences
For years, the conventional wisdom was that AI would be a double-edged sword, helping attackers find vulnerabilities, but also helping defenders close them. That framing now looks dangerously optimistic. AI is compressing attack timelines faster than most defence organisations can respond, and the gap is widest in the public sector.
Consider the baseline before AI entered the picture: the median time for an organisation to remediate half of its open, internet-facing vulnerabilities was 361 days. Exploitation, meanwhile, takes hours. One-third of exploited CVEs in the first half of 2025 showed attacker activity on or before the day of public disclosure, before most teams even knew there was anything to patch.
For federal agencies and SLED organisations, those numbers are often worse. Legacy systems, complex procurement cycles, stretched security staffs, and compliance-heavy environments all compound to slow remediation to a near halt. AI is further compressing the attacker’s window. The asymmetry is real, and it’s growing.
The early assumption was that AI, a tool equally available to attackers and defenders, would level the playing field. That assumption is proving to be incorrect. The evidence now points in one direction: in the near term, AI is likely to benefit attackers more than defenders. For the public sector, that imbalance is the problem to solve.
How government agencies can respond
Closing the gap requires action on two fronts simultaneously: hardening the software development pipeline so vulnerabilities never ship, and increasing proactive threat hunting so attackers don’t find exploitable gaps first.
-
Red team before adversaries do. The most underutilised asset in public sector security is proactive offence. Red team exercises, bug bounties, and security hackathons are no longer a nice-to-have. If AI is helping attackers discover vulnerabilities at scale, defenders need structured programs to find them first. Government agencies that invest in red teaming now will be dramatically better positioned than those waiting for CVE disclosures.
-
Shift security left, all the way to the IDE. Too many agencies still treat security as a gate at the end of the development process. With AI accelerating code production and introducing new vulnerabilities in the process, that model is untenable. Hardcoded secrets, known-vulnerable imports, and deprecated API patterns need to be flagged before a developer pushes a commit, not during an audit six months later. This maps directly to CISA’s Secure by Design principles and should be a baseline expectation for any government contractor or internal development team.
-
Enforce security policy across every pipeline, not just compliant ones. In large agencies with hundreds of active projects, inconsistent enforcement is the rule rather than the exception. Security policies need to be defined once and enforced everywhere, at the merge request level, across every group and project, with exceptions reviewed, approved, and logged. The merge request is the enforcement point. If a line of code can enter a production system without passing a defined set of controls, the gap will be found.
-
Use AI agents to discover and remediate vulnerabilities. Responding to AI-assisted attacks requires AI-assisted defence. That means deploying security agents that can scan dependency graphs across every project with full context, identify reachable vulnerable call paths, and propose fixes, all within the same governance process as human-authored changes. Leveraging AI to identify vulnerabilities during the development process ensures fewer vulnerabilities for adversaries to exploit after the software is shipped. Critically, every AI-generated fix must move through the same approval and audit trail as any other change. In a FedRAMP- or STIG-governed environment, that process discipline also makes AI remediation viable without introducing new risk.
-
Build for auditability from the start. When an auditor or oversight body asks for evidence that a specific security policy was enforced on a specific change, the answer should be immediately accessible. Evidence generated in flight, such as scan results, applied policies, approvers, and merge timestamps, serves as both a compliance and a security asset. Assembling audit evidence after the fact means discovering gaps.
The zero-day clock is already running
The window is open, but it will not stay open for long. Anthropic restricted access to its Mythos model, and its own offensive security researchers estimate that comparable attacker tooling will likely be widely available within six to twelve months. For Australian federal and state agencies, that timeline is short in procurement and delivery terms, but still long enough to take action.
This is the time to strengthen software supply chains, enforce consistent policy across pipelines, and build the proactive threat-hunting capability that modern environments now require. Agencies must reduce exposure while conditions are still shifting, and those that move now will be far more resilient.
