Startup Anthropic to Brief Regulators on 'Mythos' AI Amid Financial Cyber Risk Concerns

(London, May 18, Reuters) The Financial Times reported today, citing sources familiar with the matter, that AI startup Anthropic is expected to brief the global Financial Stability Board (FSB) on the potential for its cybersecurity AI model, Mythos, to create vulnerabilities in global financial cybersecurity. The chatbot developer was asked to present at the meeting at the request of Andrew Bailey, the Governor of the Bank of England. In addition to his role as head of the UK's central bank, Bailey also chairs the FSB, which coordinates financial regulation for the G20 members. The Financial Times noted that Anthropic will provide an explanation of its newly launched Mythos Preview AI model. Anthropic stated that Mythos, which was announced last month but has not yet been publicly released, is a new cybersecurity model designed to detect vulnerabilities that have existed for decades in current browsers, infrastructure, and software. However, cybersecurity experts warn that this advanced security system model, while capable of strengthening cybersecurity defenses, could pose an attack risk to the banking industry, which relies on older systems. The tool is described as a double-edged sword of technology: while it is a miraculous tool for patching vulnerabilities on the 'defensive side,' it can also autonomously act as a hacker with extremely destructive capabilities. In April, Bailey had warned that Mythos could pose a significant security risk to the cyber world. Reuters could not immediately verify this report. Both Anthropic and the Financial Stability Board have not yet responded to Reuters' requests for comment.