Anthropic launches Project Glasswing, an effort to prevent AI cyberattacks with AI

image via Engadget

April 7, 2026, 9:49 PM

• •Project Glasswing aims to secure critical software against AI-powered attacks.
• •Partners include Amazon Web Services, Apple, and Microsoft.
• •Uses Claude Mythos Preview to find vulnerabilities.
• •Anthropic faced a 'supply chain risk' designation from the Department of Defense.

Anthropic has launched Project Glasswing, an initiative to secure critical software against AI-powered attacks, with partners including major tech companies. The project will utilize Anthropic's unreleased Claude Mythos Preview model to identify and address vulnerabilities, aiming to prevent malicious AI use. Anthropic, known for its ethical concerns in AI, has faced challenges, including a sanction from the Department of Defense. Despite these issues, Project Glasswing represents an effort to improve cybersecurity in the AI era, though past incidents highlight the ongoing risks.

Read original article