Anthropic Launches Project Glasswing to Combat AI-Driven Cyberattacks with Major Tech Partners

Anthropic has launched Project Glasswing, a major cybersecurity initiative that brings together leading technology and infrastructure companies. The project aims to address the rising threat of AI-driven cyberattacks by using advanced artificial intelligence to secure critical software before attackers can exploit vulnerabilities.

Key Highlights

• Anthropic launches Project Glasswing with partners like Apple, Google, and Microsoft
• Claude Mythos Preview AI model has found thousands of high-severity software vulnerabilities
• Over 40 organizations have access to the Mythos system for critical infrastructure security
• Anthropic commits up to $100 million in usage credits and additional funding for open-source security
• US government officials are involved to prioritize national security in the AI era

Major Partners and Project Goals

Project Glasswing includes key partners such as Amazon Web Services, Apple, Google, Microsoft, Nvidia, and the Linux Foundation. These organizations will collaborate to use advanced AI systems for software security. The central tool is an unreleased AI model called Claude Mythos Preview.

Anthropic states that Claude Mythos Preview can identify and exploit software vulnerabilities at a level that exceeds all but the most skilled human experts. The model has already discovered thousands of high-severity vulnerabilities, including flaws in major operating systems and web browsers. This demonstrates the rapid advancement of AI tools and the potential risks if such technology is misused.

Industry Collaboration and Security Measures

Anthropic warns that the capabilities of AI models like Mythos will soon become widespread, possibly reaching actors who may not use them responsibly. Project Glasswing seeks to counter this risk by deploying powerful AI for defensive purposes.

As part of the initiative, partner organizations will use the Mythos model to scan and secure their systems. Anthropic plans to share findings with the industry to improve collective security. Over 40 organizations that develop or maintain critical software infrastructure have received access to the Mythos system.

To support the project, Anthropic is providing up to $100 million in usage credits for Mythos Preview. The company is also allocating additional funding for open-source security work. Anthropic emphasizes that no single organization can address cybersecurity challenges alone and calls for immediate action.

Government Involvement and National Security

Anthropic is working with US government officials to prioritize the security of critical infrastructure. The company notes that protecting key systems is now a central national security concern in the era of advanced AI.