Defending the Exploit Window at AI Speed

Navigating cybersecurity in an era of rapidly advancing AI requires a fresh look at how government agencies manage risk, deploy automation and execute security strategies. Frontier models like GPT-5.5-Cyber and Claude Mythos demonstrate the growing speed and sophistication of today’s threat landscape — threat actors can now weaponize vulnerabilities faster than organizations can remediate them.

During ATARC’s recent “Next-gen cyber operations: AI in action summit,” Ryan Johnson, director of solutions engineering for U.S. public sector at F5, addressed this rapidly shrinking timeframe and outlined strategies for securing the data path at AI speed.

Here are three key takeaways from his discussion.