The Federal Reserve, a key institution in global finance, did not have access to Anthropic's advanced AI model, Claude Mythos Preview, for at least three months. This left the Fed vulnerable to cybersecurity threats that other institutions were already addressing.
TL;DR
- The Federal Reserve was without access to Anthropic's Claude Mythos Preview for at least three months, despite its importance in identifying cybersecurity vulnerabilities.
- Federal Reserve Chairman Kevin Warsh testified that the Fed is still working to secure access to Mythos and other cutting-edge AI models.
- Anthropic's Mythos model is part of Project Glasswing, a cybersecurity initiative involving around 200 organizations across 15 countries.
What happened
In April, the Federal Reserve and Treasury Department held a meeting with top bank CEOs to discuss the cybersecurity threats posed by Anthropic's Claude Mythos Preview. This AI model excels at identifying software vulnerabilities and was released to select institutions as part of Project Glasswing.
Despite the Fed's critical role in financial cybersecurity, it did not have access to Mythos until at least July 15. It is unclear if the Fed has gained access since then.
Anthropic unveiled Claude Mythos Preview and Project Glasswing in early April, with around 50 organizations initially having access. This number expanded to over 150 organizations across 15 countries in June.
Why it matters
The lack of access to Mythos left the Federal Reserve vulnerable to cybersecurity threats that other institutions were addressing. This raises concerns about the Fed's ability to protect the financial system effectively.
Federal Reserve Chairman Kevin Warsh has emphasized the importance of the Fed and other institutions having access to cutting-edge AI models to identify and patch vulnerabilities. His testimony highlights the ongoing efforts to secure this access.
The rollout of Mythos has been marked by intrigue and confusion, including complications arising from Anthropic's relationship with the Trump administration and export control directives. This underscores the complex regulatory environment surrounding advanced AI models.
Key facts
- The Federal Reserve did not have access to Claude Mythos Preview as of mid-July, according to CNBC.
- Anthropic's Mythos model is part of Project Glasswing, a cybersecurity initiative involving around 200 organizations across 15 countries.
- Federal Reserve Chairman Kevin Warsh testified that the Fed is still working to secure access to Mythos and other cutting-edge AI models.
- Anthropic initially released Mythos to around 50 organizations in early April, expanding access to over 150 organizations in June.
- The Federal Reserve meeting in April was convened by former Fed leader Jerome Powell alongside Treasury Secretary Scott Bessent.
- Anthropic has been in ongoing discussions with U.S. government officials about the model, including the Cybersecurity and Infrastructure Security Agency and the Center for AI Standards and Innovation.
- The Trump administration has been taking a more active role in AI regulation, with questions remaining about who is calling the shots on AI policy behind the scenes.
Context
The Federal Reserve plays a crucial role in maintaining the stability and security of the global financial system. Its lack of access to advanced AI models like Mythos raises concerns about its ability to fulfill this role effectively.
The rollout of Mythos and the complexities surrounding its access highlight the challenges and opportunities presented by advanced AI models in the realm of cybersecurity. As AI continues to evolve, the need for robust regulatory frameworks and effective collaboration between private and public sectors becomes increasingly apparent.
The competitive landscape of AI is heating up, with Chinese models like Moonshot AI's Kimi K3 gaining traction against leading U.S. offerings. This underscores the importance of the U.S. maintaining its lead in AI innovation and regulation.
