The Lawfare Podcast Scaling Laws: Caleb Withers on the Cybersecurity Frontier in the Age of AI
Dec 5, 2025
Caleb Withers, a research associate at the Center for a New American Security, dives into the evolving landscape of cybersecurity influenced by AI. He explains how frontier AI models create advantages for attackers, increasing the scale and sophistication of cyber threats. Withers highlights persistent software vulnerabilities like SQL injection and the dual role of AI in both creating and detecting security flaws. He also emphasizes the importance of robust authentication measures and the necessity for agile policy responses to emerging risks in cyber warfare.
AI Snips
Chapters
Transcript
Episode notes
Frontier AI Is More Than Chat
- Frontier AI refers to large foundation models that are multimodal and can use computers, not just text chatbots.
- These capabilities matter in cyber because models can interact with systems and perform actions beyond language.
Running Models Changes The Cost Equation
- AI still helps defenders but frontier models change cost dynamics because bigger, longer runs keep giving returns.
- That shift could make defense more expensive and strategic choices about spending matter more.
Old Vulnerabilities Still Dominate
- Many cyber vulnerabilities are long-lived 'stubborn weaknesses' like SQL injection that persist over decades.
- Widespread low-quality code and easy tooling can proliferate these vulnerabilities even as AI empowers coders.