Anthropic has introduced its Critical Infrastructure Defense Program to give essential utility operators and security partners access to its most powerful AI models and engineers. Founding partners include major industry players like CrowdStrike, Deloitte, Accenture, and Palo Alto Networks, some of which are already using Claude to identify and patch system vulnerabilities. The initiative also features a free, automated “OSS Scanner” service designed to periodically evaluate open-source software projects and provide automated vulnerability reports. The program builds on learnings from Project Glasswing, demonstrating how rapidly AI can surface security flaws, though verifying and implementing fixes remains challenging. Key logistical questions remain regarding how participating companies will test fixes without disrupting utility operations and how computing costs will be covered.

Full report : Anthropic partners with major cybersecurity and tech firms to deploy its frontier AI models and on-site engineers to safeguard power grids and water utilities against advanced cyber threats.