- Anthropic expands Project Glasswing access to 150 organisations in over 15 countries including India
- Claude Mythos Preview helps partners identify over 10,000 critical cybersecurity vulnerabilities
- Project Glasswing involves industry, governments, and open-source groups to protect digital infrastructure
San Francisco-based AI company Anthropic has announced a major expansion of its Project Glasswing, granting access to its advanced cybersecurity model, Claude Mythos Preview, to approximately 150 new organisations across more than 15 countries. India is one of the nations whose companies will now have access to Claude Mythos Preview in a bid to bolster cybersecurity. The initiative, which began with around 50 select partners in early April, now includes organisations from Canada, Australia, France, Germany, Japan, and South Korea as well.
Project Glasswing is a collaborative effort involving partners, the cybersecurity industry, open-source software maintainers, and governments to strengthen defenses in vital digital infrastructure, including banking and financial systems.
Initial partners — such as Amazon Web Services, Microsoft, Apple, Google, Broadcom, CrowdStrike, Nvidia, Cisco, JPMorgan Chase, Palo Alto Networks, and the Linux Foundation — have already deployed Claude Mythos Preview to scan their codebases. Anthropic said its partner organisations have identified more than 10,000 high- or critical-severity vulnerabilities.
The expanded group now covers critical sectors such as power, water, healthcare, communications, and hardware, along with vendors and nonprofits whose codebases support governments and numerous other organisations globally. Anthropic intends to further expand its geographical reach in the future, it said.
Anthropic emphasised that these partners manage systems where a successful cyberattack could impact over 100 million people, posing significant risks to national and global security. The company warned that similar Mythos-class models from other AI developers could emerge within 6 to 12 months without adequate safeguards, potentially enabling misuse and more frequent and unpredictable cyberattacks.
Notably, during internal testing, Claude Mythos Preview uncovered thousands of severe vulnerabilities, including long-undetected zero-day exploits in major operating systems and browsers. While partners are using the model at scale and sharing best practices, Anthropic said it is not yet ready for general release due to challenges in developing safeguards against misuse by cyber criminals.
Also read: Meta's Own AI Helped Hackers Steal Instagram Accounts Before Emergency Patch
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.