Finance Minister Satsuki Katayama stated that the Japanese government and some of the country’s financial institutions have gained access to “Claude Mythos Preview,” the cutting-edge AI model developed by Anthropic. She made the announcement at a press conference on June 2.
Because Claude Mythos’s exceptionally high capabilities have raised concerns about misuse, Anthropic initially limited access to the U.S. government as well as some of that country’s financial institutions and tech companies. Against this backdrop, Katayama had asked U.S. Treasury Secretary Scott Bessent, who visited Japan in mid-May, for the Japanese government to be granted access to Claude Mythos. This is because cutting-edge AI poses a risk of exposing the networks of Japanese financial institutions to vulnerabilities.
The request has thus now been fulfilled. On this point, Katayama commented, “I am very pleased that the promise between our two countries was faithfully kept, and I am deeply grateful for the efforts of those on the U.S. side.”
Katayama also disclosed that “some financial institutions” in Japan had likewise gained access to Claude Mythos Preview. Asked by reporters which institutions they would be, she hinted that they were megabanks, responding: “Because there are contracts between individual companies and so on, rather than my stating it myself, I think you can all probably imagine which ones. What I will say is that they are likely of a certain (scale); after all, it is not every financial institution.”
