Payward, the parent company of Kraken, partners with Anthropic to enhance cybersecurity using AI-driven vulnerability scanning.
Overview
- Payward, the parent company of Kraken, is now involved in Anthropic’s Project Glasswing, using Claude Mythos 5 for cybersecurity.
- This collaboration aims to uncover system vulnerabilities, which will be addressed through established security measures.
- Project Glasswing offers select organizations access to advanced AI models designed to bolster cybersecurity defenses.
Payward Inc., the parent company of the cryptocurrency exchange Kraken, has partnered with Anthropic for cybersecurity enhancements via its Project Glasswing initiative. By leveraging the capabilities of Claude Mythos 5, Payward aims to pinpoint vulnerabilities within its systems.
This partnership comes on the heels of a letter signed by over 40 crypto firms, including Kraken, urging major AI labs like Anthropic and OpenAI to develop trusted-access programs for cybersecurity professionals.
In an announcement on Monday, Payward detailed its plan to utilize Mythos 5 for scanning vulnerabilities, with findings being reviewed by its security teams. Notably, Payward is the first cryptocurrency enterprise to participate in this initiative, gaining exclusive access to Claude Mythos 5.
According to Payward, this selection allows its security team to utilize advanced models in their efforts to manage complex software vulnerabilities, thereby better protecting millions of customers globally. The company also plans to notify maintainers of any vulnerabilities found in third-party open-source software, enhancing the security of its financial infrastructure and the broader cryptocurrency ecosystem.
Launched in April and expanded in June, Project Glasswing is Anthropic's initiative that equips selected organizations with its most sophisticated cyber defense models, which have already helped identify thousands of critical vulnerabilities.
This move highlights the growing recognition of AI's role in both detecting and exploiting software vulnerabilities. For instance, earlier in April, Mozilla revealed that Anthropic’s Claude Mythos had identified 271 vulnerabilities in Firefox during testing.
Last week’s coalition of cryptocurrency firms, organized by the Bitcoin Policy Institute, argued that access to cutting-edge AI is essential for developers protecting open-source financial systems from malicious actors. Payward Co-CEO Arjun Sethi commented, “Security has always been an uneven playing field. Frontier AI shifts this imbalance, enabling us to detect flaws before they can be exploited.”