A significant increase in AI-driven cyberattacks has been reported, with a staggering 72% rise year-over-year. By 2025, 87% of global organizations are expected to face such threats, raising alarms about the implications of advanced AI capabilities falling into the wrong hands.
In response to these concerns, Anthropic has launched a new initiative called Project Glasswing, enlisting over 40 companies, including major tech players like Amazon Web Services, Apple, and Google. This project aims to utilize the Claude Mythos Preview AI model to proactively identify and patch critical vulnerabilities before malicious actors can exploit them.
The Claude Mythos Preview model has already uncovered thousands of vulnerabilities, including significant flaws in every major operating system and web browser. Notably, it discovered a 27-year-old bug in OpenBSD and a 16-year-old flaw in the FFmpeg library, showcasing the potential of AI in enhancing cybersecurity measures.
As Cybersecurity stocks reacted negatively to the unveiling of Claude Mythos, Anthropic emphasized the importance of collaboration in addressing these vulnerabilities, which are often subtle and challenging to detect, and can be decades old.