By Joe Levy, Chief Executive Officer, Sophos
The rapid evolution of agentic AI, the kind that can act independently, make decisions, and chain thoughts together to seek goals, calls for a shift in how we think about cybersecurity. This is no longer just about models that generate content. This is the real-time emergence of systems that behave more like employees. Not mere technology or tools, but something else entirely.
At this year’s US Cybersecurity Group Summer Meeting, led by Aspen Digital, I provided opening statements for a session of cybersecurity experts exploring how rising agency in AI is reshaping both cyber offense and defense capabilities. I wanted to share some of my personal views here to recap my thoughts and continue the conversation beyond the forum walls.
As we strengthen cyber defenses for all organizations, we need to imagine the potential impact of current and future generations of artificial intelligence operating agentically.
These systems won’t just assist with security tasks; they’ll autonomously perform a wide range of actions traditionally handled by humans, on both the offensive and defensive sides. To navigate this future responsibly, we need to understand the benefits, risks, and challenges it will introduce to how we build, develop, and trust our IT operations and cyber programs.
This piece is part of an Aspen Digital series of perspectives on the evolving space of intergovernmental cyber policy, including challenges and best practices for building state, local, tribal and territorial capacity and how governments can collaborate effectively.
The views represented herein are those of the author(s) and do not necessarily reflect the views of the Aspen Institute, its programs, staff, volunteers, participants, or its trustees.