Verizon Joins Anthropic AI Security Venture: Project Glasswing

Verizon is now the first telecom operator to join Project Glasswing, an AI-driven bug hunting taskforce. This project was created by AI vendor Anthropic. Verizon states this move will help them find and fix complex vulnerabilities, keeping their network secure. Other members of Project Glasswing include Cisco, Nvidia, Broadcom, and Palo Alto Networks. Project Glasswing, launched in April, uses Anthropic's new large language model, Claude Mythos Preview. Anthropic says Mythos Preview is good at cybersecurity tasks, having found thousands of high-severity vulnerabilities without human help. This Mythos model is not publicly available; its use is limited to Glasswing partners and over 40 other organizations. Verizon's CEO, Dan Schulman, noted that Verizon is the only telecommunications company using this model. He emphasized that customers depend on their network's security. Verizon says it follows strict policies to ensure all new technologies are evaluated safely. This matters because AI is being used by both companies and cyber attackers, making AI-driven security responses increasingly important.