Security companies are accelerating efforts to counter threats that exploit AI models, with momentum building around agentic AI-based security. Zscaler introduced two security products for AI agents and an AI Access Graph, while Check Point launched agent-based exposure validation. Concerns are rising over a “patch gap” as AI can find vulnerabilities and write attack code within hours. Companies including Rubrik, Samsung SDS and others also announced new products and partnerships.

Anthropic is sending engineers to the U.S. National Security Agency to help it use the Mithos AI model for offensive cyber operations, the Financial Times reported, citing two sources. The report said about 6 employees are deployed on site to guide use of the technology and customise the model for specific purposes. It is unclear whether they are involved in actual operations.