Facing the looming threat of a powerful new AI model from Anthropic, tech giants have banded together in an unprecedented industry consortium. Dubbed Project Glasswing, this collaboration aims to fortify cybersecurity against a future where AI capabilities could redefine the landscape.

The group includes major players like Microsoft, Apple and Google, working collectively on the Mythos Preview model. It’s not just about locking down current vulnerabilities; these companies are also exploring how to adapt existing security practices in anticipation of more advanced models in the coming years.

Anthropic CEO Dario Amodei highlighted the broader implications: ‘More powerful models are going to come from us and others, so we do need a plan to respond to this.’ The model’s capabilities extend beyond simple vulnerability discovery, encompassing everything from producing attack chains to assessing software security without access to source code.

The staggered release approach mirrors principles of coordinated vulnerability disclosure. But the real challenge lies in how these AI tools are used. As Graham puts it, ‘Done not carefully, this could be a meaningfully accelerant for attackers.’

Ultimately, Project Glasswing represents an attempt to harness AI for good—a collaborative effort that acknowledges the inevitable and seeks to prepare humanity for what’s to come.