Watch: From tool to actor — the rise of agentic AI
The growth of Agentic AI is reshaping the threat landscape. AI agents — such as hackerbot-claw — can execute sophisticated supply chain attacks quickly, and that requires a fundamental rethink of how we engineer trust in modern systems.
We need to adapt our approach to security. The oft-used phrase “human-in-the-loop” collapses under high-throughput requirements.
The human role must shift from task checkers to system governors, using exception-based intervention or “cognitive speedbumps” when confidence boundaries are breached.
Delegation-with-authority to AI workloads requires rigorous application of security principles, clarity of expectations and clear authority boundaries.
The presentation was delivered at CyNam’s summer headline event, ‘In Machines We Trust.’
Presenter:
Steve West, Secure Government & Critical Infrastructure Director at Zaizi, is a technologist with a career spanning both the private and public sectors.
He spent over 10 years working in national security, focusing on modern architectures and emerging technologies. During this time, he held a CTO position with the Foreign and Commonwealth Office and was a founding member of the National Security Strategic Investment Fund.
