Anthropic joins hands with AWS, Apple, Microsoft for Project Glasswing

Initiative aims to deploy Claude Mythos AI for defensive cybersecurity, with partners working to identify vulnerabilities and strengthen global digital infrastructure

Shortly after releasing a preview of its new general-purpose large language model (LLM), Claude Mythos, Anthropic has joined forces with several leading companies to launch Project Glasswing, an initiative to address potential misuse of its LLM’s coding capabilities.

 

The new initiative will see leading global companies such as Amazon Web Services (AWS), Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, Nvidia, and Palo Alto Networks join hands with the Dario Amodei-led Anthropic to find solutions to deploy Claude Mythos’s prowess “for defensive purposes”.

 

“Mythos Preview has already found thousands of high-severity vulnerabilities, including some in every major operating system and web browser. Given the rate of AI progress, it will not be long before such capabilities proliferate, potentially beyond actors who are committed to deploying them safely. The fallout — for economies, public safety, and national security — could be severe,” Anthropic said in a blog post announcing Project Glasswing.

 

As part of this project, Anthropic and its partners will use Claude Mythos for their defensive security work and share the learnings with other industry partners worldwide.

 

Further, companies that have joined this project will also have access to the LLM’s preview to find and fix cybersecurity vulnerabilities and weaknesses in their own foundational systems.

 

“We anticipate this work will focus on tasks like local vulnerability detection, black box testing of binaries, securing endpoints, and penetration testing of systems,” Anthropic said.

 

While the project partners will share their learnings with each other as much as possible, Anthropic will publicly release the findings within 90 days, which will include details on aspects such as vulnerability disclosure, software update, open-source and supply-chain security, software development lifecycle and secure-by-design practices, standards for regulated industries, triage scaling and automation, and patching automation.

 

“We will also collaborate with leading security organisations to produce a set of practical recommendations for how security practices should evolve in the AI era,” the company said.

 

Though Anthropic does not intend to make Claude Mythos Preview general available to everyone, the company said it wants to enable its users to “safely deploy Mythos-class models at scale — for cybersecurity purposes, but also for the myriad other benefits that such highly capable models will bring”.