Anthropic’s New “Project Glasswing” Will Use AI to Fight Against AI Cyber Attacks

Anthropic has introduced a new initiative called Project Glasswing, aimed at protecting critical software systems from AI-driven cyberattacks.

The project brings together major technology and security players, including Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks.

Defending Critical Systems

Project Glasswing is designed to address concerns about how artificial intelligence could introduce new cybersecurity risks. The initiative focuses on securing widely used software systems against attacks powered by advanced AI tools.

Participants will use Claude Mythos Preview, an unreleased general-purpose model developed by Anthropic, to strengthen their own security efforts.

AI Model Identifies Vulnerabilities

According to the company, Claude Mythos Preview has already identified thousands of exploitable vulnerabilities, including issues across major operating systems and web browsers.

Anthropic said the goal is to apply its AI tools defensively to prevent misuse that could lead to broader economic and security risks.

Anthropic has been among the companies raising concerns about the ethical use of artificial intelligence. Earlier this year, it declined to remove safeguards from its systems for use by the Pentagon. That decision led to the US Department of Defense assigning Anthropic a “supply chain risk” designation.

Despite its focus on security, Anthropic’s technology has already been linked to misuse. Reports indicate that the Claude model was used in a cyberattack targeting multiple government agencies in Mexico in February.

The launch of Project Glasswing comes as companies and governments continue to assess how to manage the risks associated with increasingly capable AI systems.