Emerging AI Capabilities and the Future of Cybersecurity
From CNAS Live by Center for a New American Security | CNAS
December 18, 2025 · 59 min · Episode 7
About this episode
The episode discusses the impact of emerging AI capabilities on cybersecurity, featuring insights from a panel of experts.
On November 10, CNAS hosted a panel on recent advances in frontier artificial intelligence (AI) models and their potential to transform cyber offense and defense alike. Featuring the new report, Tipping the Scales: Emerging AI Capabilities and the Cyber Offense-Defense Balance, CNAS Technology and National Security Program expert Caleb Withers examines how capable today’s AI systems are at cyber-relevant tasks, which emerging capabilities matter most, and what defenders, policymakers, and industry should do to prepare. The discussion drew on those insights to cover a range of topics from recent advances in frontier AI intelligence models to the attacker-defender imbalance in cybersecurity operations, and more. The panel featured: Moderator Derek B. Johnson, journalist, CyberScoop Caleb Withers, Fellow, Center for a New American Security Andrew Lohn, Senior Fellow, Georgetown’s Center for Security and Emerging Technology Chris Rohlf, Security Engineer, Meta To learn more about upcoming events at CNAS, head to cnas.org/events.
People in this episode
Host: Derek B. Johnson
Guests: Caleb Withers, Andrew Lohn, Chris Rohlf
Topics covered
- artificial intelligence
- cybersecurity
- cyber offense
- cyber defense
- emerging technologies
- policy preparation
Keywords
- AI systems
- cyber-relevant tasks
- defenders
- policymakers
- industry preparation
- attacker-defender imbalance
- frontier AI models
Mentioned in this episode
Organizations: Center for a New American Security, Georgetown’s Center for Security and Emerging Technology, Meta, CyberScoop
Books & works: Tipping the Scales: Emerging AI Capabilities and the Cyber Offense-Defense Balance
Explore listener stats, chart rankings, contacts and more on the CNAS Live podcast page.