Japan's three largest banks are stepping up cybersecurity measures following warnings that Anthropic's Claude Mythos artificial intelligence model could expose vulnerabilities in financial systems, Nikkei reported Friday.
The banks are expected to gain access to the model as early as this month to help identify software flaws that developers may have overlooked, according to the report.
Mitsubishi UFJ Financial Group (TYO:8306) plans to establish a dedicated cybersecurity team, while Mizuho Financial Group (TYO:8411) has formed a cyberattack task force. Sumitomo Mitsui Financial Group (TYO:8316) has also created a working group to review systems across the company, the report said.
The International Monetary Fund warned this month that cyberattacks using advanced AI models such as Mythos could threaten financial stability if multiple institutions are targeted simultaneously, potentially disrupting settlements and triggering liquidity stress, according to the report.
Japan's Financial Services Agency has launched a public-private working group involving major banks, technology companies and industry bodies to discuss countermeasures, including stronger management involvement and increased cybersecurity investment, the report said.
(Market Chatter news is derived from conversations with market professionals globally. This information is believed to be from reliable sources but may include rumor and speculation. Accuracy is not guaranteed.)