**Anthropic Plans to Provide Access, Observations of Spread in Cybersecurity Sector**
In light of ongoing controversy over the potential risks associated with Anthropic’s high-performance AI model ‘Claude Mythos,’ it is being reported that European banks are considering adopting the model. Industry observers suggest that banks’ competition to integrate AI could spread beyond mere operational efficiency to encompass the entire cybersecurity framework.
According to a report by Reuters on the 21st (local time), Anthropic is planning to provide access to the Mythos model to banks in Europe and the UK in the near future. Currently, the model is being tested in approximately 40 key infrastructure-related institutions, including major U.S. financial firms like JPMorgan Chase and Bank of America (BoA).
Mythos is evaluated as an AI model equipped with complex coding tasks and system analysis capabilities. However, as the model shows excellent performance in detecting software security vulnerabilities, concerns are rising that it could be misused for crimes such as hacking, potentially exposing the entire financial system to risk. Analysts suggest that Mythos could become a ‘security nuclear bomb’ for outdated financial systems.
In response to the fears stemming from Mythos, major institutions worldwide are hastily assessing the potential risks. On the 12th, it was reported that the Bank of England (BOE), the Financial Conduct Authority (FCA), and the Treasury met with the National Cyber Security Centre (NCSC) and major banks to evaluate the negative impact Mythos might have on the financial system. It also emerged as a major agenda item at the International Monetary Fund (IMF) and World Bank Spring Meetings, with finance ministers and central bank governors expressing unanimous concern.
European monetary authorities are also raising their alert levels. Joachim Nagel, the President of the Bundesbank (German Central Bank), warned that “Mythos can quickly find and exploit security vulnerabilities in financial institutions’ software,” emphasizing the need to prevent misuse as AI utilization in the financial sector could heighten new cyber risks.
The European Central Bank (ECB) has also initiated a response, announcing on the 15th its intention to assess the risks related to Mythos among banks. On the 20th, it reported that international banks and regulatory authorities are urgently assessing the impact of Mythos. Financial authorities in Europe, the UK, and Asia are closely monitoring the possibility of Mythos surpassing existing risk management frameworks.
Despite these concerns, the urgency among banks to adopt Mythos is driven by the need to enhance productivity and security capabilities. High-performance AI is seen as highly effective in handling vast amounts of internal documents, code review, regulatory responses, and security vulnerability detection. It is especially noted for its potential to impact not just process automation but also financial security and system stability.
For the domestic financial sector, the move by European banks to adopt Mythos amidst a lowering AI threshold could exert considerable pressure. An industry insider commented, “As global financial firms accelerate AI-based innovation, the domestic financial sector must also enhance its AI adoption speed and risk management capabilities to maintain competitiveness.”