Loading story…
Organizations and researchers reported that AI is materially reshaping cybersecurity operations, both by increasing the scale of threats defenders must process and by expanding what automated systems can do. Commonwealth Bank said it built its own agentic AI threat-hunting capability because commercial tools were not keeping pace with emerging AI-enabled threats, citing a jump from tens of millions of signals to hundreds of billions and describing repeated phishing infrastructure that appeared to reuse common backend code, sometimes with artifacts associated with AI coding tools. The bank said it is using internally built AI to ingest threat intelligence, analyze it against its own environment, and help junior analysts operate with access to knowledge normally held by senior staff.
Separate reporting and research reinforced the broader operational and strategic impact of AI on security. TechTarget outlined how security leaders are being pushed to justify AI investment through measurable gains in analyst throughput, risk reduction, and cost avoidance, including improvements in MTTD, MTTR, vulnerability remediation, and configuration management coverage. The UK AI Security Institute published testing showing that frontier AI agents are becoming more capable in multi-step cyber-attack scenarios on simulated network ranges, with newer models completing substantially more attack steps and benefiting further from increased inference-time compute. Taken together, the material is not fluff: it documents enterprise deployment, board-level investment considerations, and government-backed evidence that AI systems are improving at complex offensive cyber tasks.
Track how attackers are adapting to this technology.
5 events from the most recent confirmed update back to the earliest known activity.
Commonwealth Bank of Australia said it developed its own agentic AI tools for cyber defense because commercial security vendors were not keeping pace with AI-enabled threats. The bank built one agent to correlate threat intelligence with internal environments and another to search for indicators of compromise and generate reports, cutting some threat assessment work from two days to 30 minutes.
TechTarget published an article focused on calculating the return on investment of AI in cybersecurity. No specific discrete real-world incident, breach, or operational event was described in the provided reference.
The UK AI Security Institute published research testing seven frontier AI models in realistic multi-step cyberattack scenarios across a corporate network and an industrial control system range. The study found rapid gains in offensive capability across model generations and with higher inference-time compute, while also noting current limitations in complex ICS tasks and patching an unintended attack path discovered during testing.
Researchers proposed an environment-grounded multi-agent architecture for autonomous penetration testing in robotics environments, using shared graph-based memory to track topology, communications, vulnerabilities, and exploit attempts. In a ROS/ROS2 robotics Capture-the-Flag evaluation, the system reportedly solved the challenge in all five runs while emphasizing traceability and human oversight.
Researchers introduced CTI-REALM, a benchmark for evaluating AI agents on cyber threat intelligence interpretation and security detection rule development using emulated attacks across Linux, cloud, and AKS environments. The evaluation reported results for 16 frontier models, finding Claude Opus 4.6 (High) achieved the highest overall reward and that CTI-specific tools and memory augmentation improved performance.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
5 references tracked. Mallory keeps watching after this page renders.
go.theregister.com
Open sourcetechtarget.com
Open sourceaisi.gov.uk
Open sourceinfosec.pub
Open sourceinfosec.pub
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Artificial intelligence is rapidly reshaping the cybersecurity landscape, enabling both attackers and defenders to operate with unprecedented speed and sophistication. Security leaders and experts warn that AI-driven malware, automated spear-phishing, and adaptive attack campaigns are already outpacing traditional defenses, as highlighted in recent Congressional hearings and industry research. Notably, Google's threat intelligence team has observed adversaries leveraging large language models to generate malicious scripts and obfuscate code, while researchers have documented the first advanced, AI-enabled cyber-espionage campaigns attributed to nation-state actors. At the same time, AI is being used to automate vulnerability discovery, with new agents like ARTEMIS outperforming most human penetration testers in live enterprise environments, and academic teams developing AI systems capable of autonomously defending wireless networks from jamming attacks. The dual-edged nature of AI is also driving a widening gap between organizations able to invest in advanced security and those falling below the 'security poverty line.' Predictions for 2026 emphasize that AI will lower the barrier for attackers while raising the cost and complexity of effective defense, forcing security and business leaders to rethink resilience strategies. The use of AI in both offensive and defensive operations is fundamentally altering the economics, speed, and scale of cyber conflict, making continuous adaptation and investment in AI-driven security capabilities a strategic imperative for organizations worldwide.
Jun 29, 2026
The rapid advancement and democratization of artificial intelligence have fundamentally altered the cybersecurity landscape, enabling both defenders and attackers to operate with unprecedented speed and sophistication. Security researchers have demonstrated that large language models can generate fully functional ransomware in under 30 seconds, drastically lowering the barrier for threat actors to create and iterate on malicious code. While some AI models still fail to produce working exploits, a significant portion succeed, raising concerns about the ease with which attackers can leverage these tools. At the same time, organizations are increasingly relying on AI for threat detection, analytics, and intrusion analysis, with many security leaders viewing AI as a necessary force multiplier to address skill shortages and burnout within their teams. Despite the promise of AI-driven defense, the technology introduces new risks, as evidenced by reports of cyber incidents linked to AI tools and concerns that automation may erode human decision-making. Industry surveys reveal that a majority of cybersecurity executives feel overwhelmed by threats without AI, yet remain wary of overreliance. Looking ahead, AI-powered defense systems are expected to become even more autonomous and adaptive, reducing incident response times and reshaping the strategic priorities of enterprises and governments alike. The evolving interplay between AI-enabled attacks and defenses underscores the urgent need for scalable prevention strategies and a renewed focus on digital trust in an increasingly automated world.
Jun 29, 2026
Security leaders across industries are increasingly concerned about the rapid evolution of AI-enabled cyberattacks, which are now among the top threats facing enterprises. Recent research highlights that cybercriminals are leveraging artificial intelligence to automate and enhance attack chains, including the use of deepfakes, automated phishing, and AI-generated malware. These AI-driven threats are capable of executing full attack sequences autonomously, from reconnaissance to data exfiltration, at speeds and scales previously unattainable by human operators. Security teams are responding by investing heavily in AI-powered defensive tools, aiming to accelerate detection, triage, and containment of threats. However, experts caution that AI should be used as a 'copilot' rather than an 'autopilot,' emphasizing the necessity of human oversight to ensure effective and responsible use of these technologies. The human element remains a critical vulnerability, as attackers use generative AI to craft highly convincing social engineering campaigns, including synthetic audio and video, which can bypass traditional awareness programs. The arms race between offensive and defensive AI is intensifying, with both sides seeking to outpace the other in sophistication and automation. Security leaders are also grappling with the challenge of integrating AI into their broader risk management and governance frameworks, ensuring that AI-driven solutions align with organizational policies and regulatory requirements. The expanding role of the CISO now includes oversight of AI risk, reflecting the technology's growing impact on enterprise security posture. As AI becomes more embedded in both attack and defense, organizations are re-evaluating their incident response strategies, workforce training, and investment priorities. The shift towards AI-driven security operations is not without challenges, including the risk of over-reliance on automation and the need for continuous adaptation to evolving threat tactics. Industry studies indicate that while AI can handle routine security tasks, complex and strategic decision-making still requires skilled human analysts. The ongoing development of AI in cybersecurity is reshaping the landscape, demanding new approaches to both technology deployment and leadership. Security teams are urged to balance innovation with caution, ensuring that AI augments rather than replaces critical human judgment. The future of enterprise security will likely be defined by the effectiveness of this human-AI partnership in countering increasingly sophisticated, AI-powered adversaries.
Jun 29, 2026