Agentic attacks demand agentic defenses (www.techradar.com)

Anthropic has exposed the first reported AI cyber espionage campaign, highlighting how quickly threat actors are adopting agentic AI to dominate cyberattacks. This development is alarming for the AI/ML community as it indicates that malicious operators can now leverage sophisticated AI techniques to automate multi-stage assaults with minimal human oversight. The report reveals that adversaries like North Korean and Chinese actors are using AI to autonomously conduct up to 90% of their operational tasks, including vulnerability exploitation, with little human intervention needed. In response to this threat escalation, the UK's Cyber Action Plan aims to bolster defenses against these advanced attacks. Security teams are urged to adopt agentic defenses that mirror offensive strategies by employing their own AI agents. These agents can autonomously manage vulnerability triage and remediation processes, allowing organizations to respond faster and mitigate risks in real-time. Integrating knowledge graphs facilitates contextual awareness, empowering defenders to prioritize risks and efficiently deploy human oversight only when necessary. As the balance between offensive tactics and defensive capabilities shifts, the need for equally sophisticated AI defenses becomes critical to safeguarding against the rapid escalation of cyber threats.