AI startup Anthropic has introduced Project Glasswing, a cybersecurity initiative that brings together major technology firms and critical infrastructure organisations to test an advanced, unreleased AI model designed to identify and help address software vulnerabilities at scale.
At the centre of the project is a frontier AI system with strong coding and reasoning capabilities, which the company claims can analyse high-risk flaws in operating systems, web browsers and widely used software. Early testing has reportedly uncovered numerous serious vulnerabilities, including some that had gone undetected for years.
Unlike typical AI releases, the model – referred to by partners as Claude Mythos Preview – will not be made publicly available due to concerns over potential misuse. Instead, access is limited to a closed group of collaborators across the tech and security ecosystem.
Microsoft, one of the participants, said the initiative marks a shift in how cybersecurity is approached at scale. “As we enter a phase where cybersecurity is no longer bound by purely human capacity, the opportunity to use AI responsibly to improve security and reduce risk at scale is unprecedented,” said Igor Tsyganskiy, Global CISO and EVP of Security and Microsoft Research at Microsoft.
“Joining Project Glasswing… allows us to identify and mitigate risk early… When tested against CTI-REALM, our open-source security benchmark, Claude Mythos Preview showed substantial improvements compared to previous models.”
Amazon Web Services said it has already begun applying the model in its internal systems. “At AWS, we build defenses before threats emerge… AI is central to our ability to defend at scale,” said Amy Herzog, Vice President and CISO at Amazon Web Services.
“We’ve been testing Claude Mythos Preview in our own security operations… where it’s already helping us strengthen our code.”
Story continues below this ad
On the other hand, Google emphasised the importance of industry-wide collaboration in tackling emerging threats.“Google is pleased to see this cross-industry cybersecurity initiative coming together… It’s always been critical that the industry work together on emerging security issues,” said Heather Adkins, VP of Security Engineering at Google.
“We have long believed that AI poses new challenges and opens new opportunities in cyber defense.”
Anthropic said the goal of Project Glasswing is to strengthen defensive cybersecurity capabilities in the face of increasingly sophisticated AI-driven threats. The model can identify vulnerabilities and, in some cases, assist in understanding how they could be exploited. This demonstrates its defensive benefits and potential risks.
The initiative is also positioned as a way to improve security in open-source software, which underpins much of today’s digital infrastructure but often lacks dedicated resources. By providing access to advanced AI tools, Anthropic aims to help maintainers proactively detect and patch vulnerabilities.
Story continues below this ad
The company has committed up to $100 million in usage credits and $4 million toward open-source security efforts, and plans to expand participation to more organisations responsible for critical systems.
The launch comes amid growing concern over the rise of AI-powered cyberattacks, with experts warning that increasingly capable models could accelerate both vulnerability discovery and exploitation. Anthropic argues Project Glasswing is an early effort to ensure such capabilities are deployed responsibly and primarily for defensive purposes.
Contact to : xlf550402@gmail.com
Copyright © boyuanhulian 2020 - 2023. All Right Reserved.