Key facts
- Top Japanese banks have regained access to Anthropic's Claude Mythos AI model.
- Access was previously suspended by the U.S. government due to cybersecurity concerns.
- The AI model has an unprecedented ability to detect software vulnerabilities.
- Initial access was limited to a few U.S. companies and organizations, including JPMorgan Chase.
- The U.S. Treasury Secretary's visit to Tokyo was linked to the renewed access for Japanese banks.
Top Japanese banks, including MUFG Bank, Sumitomo Mitsui Banking Corp., and Mizuho Bank, are set to regain access to Anthropic's powerful Claude Mythos artificial intelligence model. This comes after the U.S. government had ordered a suspension of overseas use in June due to concerns about the AI's ability to identify cybersecurity vulnerabilities.
The AI model, developed by Anthropic, has an unprecedented capacity to detect software flaws, raising fears that malicious actors could exploit it to disrupt critical infrastructure. Consequently, access has been highly restricted since its release, with only a select few U.S. companies and organizations, such as JPMorgan Chase as part of its Project Glasswing initiative, initially granted permission.
The renewed access for Japanese megabanks was reportedly discussed during a recent visit to Tokyo by U.S. Treasury Secretary Scott Bessent. While Sumitomo Mitsui Financial Group's CEO Toru Nakashima stated he had no specific details on access rights, he confirmed the bank has established an internal working group to focus on Mythos.
The widespread concern over Mythos's potential risks has paradoxically fueled optimism about Anthropic, a firm founded by former OpenAI employees. The company is reportedly in discussions to raise funds at a valuation exceeding $900 billion.