Anthropic's cutting-edge AI models are entering a new phase of deployment? It is rumored that Japan's three major banks will be granted access to Mythos.

According to informed sources, after the limited release of the Anthropic artificial intelligence (AI) model Mythos last month, which triggered a new round of cybersecurity concerns, the three major banking giants in Japan are set to receive access to the model. The sources indicate that Mitsubishi UFJ Financial Group (MUFG.US), Sumitomo Mitsui Financial Group (SMFG.US), and Mizuho Financial Group (MFG.US) are all expected to gain access to the AI model. Reports suggest that the three major banks in Japan may receive access to Mythos by the end of May, and the relevant banks were informed of this during a recent visit to Tokyo by U.S. Treasury Secretary Janet Yellen. This move comes at a time when global financial and regulatory institutions are becoming increasingly vigilant about the risks posed by Mythos. The initial release of Mythos by Anthropic is not just an ordinary AI chatbot, but a preview version of an advanced general large model that has not been fully released. Its most outstanding capabilities focus on high-level code understanding, vulnerability discovery, and vulnerability exploit generation. Anthropic officially disclosed that the Mythos Preview can discover and exploit all types of "zero-day vulnerabilities" in popular operating systems and browsers during testing, and in some cases, it can autonomously generate complex exploit programs with minimal human intervention. Zero-day vulnerabilities refer to security flaws in software, hardware, or firmware that were previously unknown and for which there is no patch to fix. Once discovered and immediately exploited, the defending side has little to no preparation time, making it one of the most dangerous aspects of cybersecurity. The reason why global financial regulatory institutions, including the Federal Reserve, are nervous is that Mythos' powerful ability to exploit "zero-day vulnerabilities" means that the IT risks for financial systems are no longer limited to single institutions being attacked. If the ability of AI large models like Mythos to discover, weaponize, and exploit "zero-day vulnerabilities" significantly lowers the threshold, it could lead to major problems such as account theft, payment interruptions, critical database exposure, or even the destruction of financial stability. Due to the widespread operation of complex legacy systems in the banking industry with overlapping new and old technologies, highly interconnected systems, and broad attack surfaces, models like Mythos excel at identifying vulnerabilities in such complex code and protocol environments and deducing exploitation paths. If Japan's major banks gain access, it will signify a new phase in the deployment of the AI model by Anthropic. Thus far, the vast majority of global Financial Institutions, Inc. have not been able to use Mythos. Currently, access to Mythos is limited to a few American companies and institutions. JPMorgan Chase was the first bank to receive access to Mythos as part of Anthropic's "Project Glasswing." This project is an AI cybersecurity initiative launched by Anthropic on April 7, aimed at using leading technology companies to actively discover and repair critical software security vulnerabilities using their cutting-edge AI model Mythos. Toru Nakashima, CEO of Sumitomo Mitsui Financial Group, stated on Wednesday that he had not heard about access permissions and therefore was not aware of specific details. He added that the bank has established an internal working group to study Mythos-related issues specifically. In addition, the Japanese Financial Services Agency plans to hold its first joint public-private sector working group meeting on Thursday to enhance cybersecurity measures. Participants are expected to include the three major banks in Japan and other lending institutions, as well as companies such as NTT Data Group Corp. and Nomura Research Institute. Officials from the Japanese Ministry of Finance and the Bank of Japan are also expected to attend the meeting. Global regulatory agencies keep a close eye on Mythos Germany's banking regulatory agency warned on Tuesday that due to advancements in artificial intelligence (AI), cybersecurity risks are "constantly rising" and "very significant," and announced the establishment of a new department to conduct targeted inspections on Financial Institutions, Inc. Mark Branson, Chairman of the Federal Financial Supervisory Authority (BaFin), stated, "These new AI models can rapidly identify a large number of vulnerabilities in both old and new IT systems. They will also be able to exploit the vulnerabilities they discover at an increasingly fast pace." Prior to this warning issued by Germany's banking regulatory agency on Tuesday, several of the highest-level financial regulatory agencies worldwide had expressed concerns about the cybersecurity risks that Anthropic's cutting-edge AI model Mythos poses to the banking industry, and have been intensifying their assessment of Financial Institutions, Inc.'s preparedness to respond to such risks. Some senior technical experts focused on the financial industry have suggested that this AI large model may have the ability to be used for large-scale disruption of the stability of banking systems. Experts have noted that Mythos' strong capabilities in high-level agent programming give it a potential unprecedented ability to identify vulnerabilities in cybersecurity systems, prompting global policymakers and financial system regulators, such as the Federal Reserve and the Bank of England, to scrutinize the development trajectory of Mythos more closely. These regulatory agencies have already regarded Mythos as a potentially central new source of risks that could impact banking systems, payment clearing, and the security of critical financial databases.