Making AI Work for Cyber Defenders: A Strategy for Strengthening U.S. Cybersecurity

Center for Strategic and International Studies  |  Carol Kuntz, Lauryn Williams

Frontier artificial intelligence (AI) models with advanced cyber capabilities, such as Anthropic’s Mythos and OpenAI’s GPT-5.5-Cyber, have emerged, leaving little time to address critical vulnerabilities in U.S. cyber defenses. These advanced models can identify and exploit previously unseen software flaws at machine speed, threatening sensitive government and civilian networks immediately.

To counter this rapid technological shift, the United States must transition from traditional human-speed workflows to an active, AI-embedded defensive posture. This transition requires the executive branch and Congress to incentivize cross-sector collaboration, prioritizing the patching of fragile infrastructure within municipal utilities, small businesses, and open-source communities. Additionally, the federal government must establish robust pre- and post-release review frameworks for frontier models to evaluate dangerous offensive capabilities and misaligned behaviors. Ultimately, securing the nation's digital ecosystem depends on constructing resilient public-private institutions capable of continuously monitoring threats and safely deploying machine-speed countermeasures before strategic competitors like China can exploit them.

Comment

Automated cyber patching introduces severe operational risks for military command networks. Machine-speed updates can inadvertently disrupt critical command and control systems during active operations. The US Cyber Command's Defend Forward doctrine requires human-in-the-loop verification for weapon systems. India's Defence Cyber Agency must adopt similar safeguards for its operational networks.

To read in details click here