TOKYO, Japan: Japan’s three megabanks have gained access to Google LLC’s new artificial intelligence (AI) model with high capability to detect vulnerabilities in computer systems, Japanese Finance Minister, Satsuki Katayama, said today, Kyodo News reported.
Amid the rising threats to financial systems, MUFG Bank, Sumitomo Mitsui Banking Corp, and Mizuho Bank were granted rights to use sophisticated AI models developed by United States (US) startup, Anthropic, and ChatGPT developer, OpenAI.
Concerns have been raised that such models could be exploited to attack financial institutions and other organisations.
Google’s CodeMender model not only detects vulnerabilities but can automatically apply security patches, Katayama told reporters after her meeting with Hiroshi Lockheimer, Chief Product Officer at Google parent, Alphabet Inc, at the Finance Ministry.
“The trends in technology are changing so rapidly, what is considered the best keeps changing,” Katayama said. “We want to make smart choices that will serve our national interest.”
Katayama also said she was told by US Treasury Secretary, Scott Bessent, during an online meeting this week that Japanese financial institutions would soon be able to access Anthropic’s Claude Mythos model again.
The institutions’ access to Claude Mythos has been suspended following a US Government directive due to national security concerns. – BERNAMA-KYODO
