Anthropic, “Mythos Preview has already identified thousands of critical vulnerabilities, with some present in every leading operating system and web browser.” This is alarming news following a recent report from Google highlighting AI as a potential hacker super weapon.
Nonetheless, Anthropic’s Mythos AI will not be launched until the company feels assured that the world is prepared for the cybersecurity challenges it may bring. Under an initiative named Project Glasswing, Anthropic is collaborating with many of the most significant tech firms “to secure the world’s vital software.” The participants in Project Glasswing include Apple, Amazon Web Services, Google, Microsoft, and other prominent entities.
Mythos is showcasing such proficiency in coding that it can effortlessly uncover hidden zero-day vulnerabilities in codebases previously considered secure. Currently, Anthropic’s collaborators in Project Glasswing are utilizing Mythos Preview for protection, while specialists continue their investigations and exchange their insights across the sector. Nevertheless, there remains a very tangible chance that cybercriminals will soon exploit Claude Mythos and similarly potent models to undertake extraordinary hacking attempts in the near future.