Japan And United States Financial Sector: Press Conference By KATAYAMA Satsuki, Japan Minister Of Finance And Minister Of State For Financial Services

I would like to ask about Mythos. Over the weekend, it was announced that Mythos would be suspended. Minister, I believe you posted on X that there has been no change in the current understanding between the Ministries of Finance of Japan and the United States. Currently, the Japanese government and some financial institutions are able to use this technology. How do you assess the potential impact on Japan going forward?

Since this issue first arose in April, we have been considering how we—being responsible for the financial sector, which is at the forefront of defense and requires the highest level of protection—should respond, with a focus on enabling testing to begin as quickly as possible. As a result, compared to other countries, testing was able to start relatively early in Japan, and we are currently in the process of establishing a testing environment. Since we had already been in the process of starting these discussions, whether or not this process will be suspended is, at this stage, not necessarily a major issue. Rather, as we have already raised with the U.S., there have been concerns that the global release of Mythos 5 and Fable 5, even for just three days, may have increased the associated risks. At this point, however, we have not yet received a precise and official explanation from the U.S. side regarding this matter. This matter is expected to be discussed at the G7 Summit on the 17th, and we have also been exchanging views locally in Evian. That said, we believe that some form of explanation should, of course, be provided by the U.S., which is the party concerned. In terms of the possibilities, it is quite possible that the level of risk may have increased, and on that point, today, in fact, the Chief Technology Officer of OpenAI is in Japan, as well as Mr. Son, representing a partner that has invested in OpenAI, and who is also involved in a project to develop defensive AI together with OpenAI here in Japan. Taking such opportunities into account, as the Japanese authorities, we intend to clearly convey what needs to be said to the U.S. If the U.S. has some authority over what information is disclosed and how it is disclosed, then it would be sufficient for us to be treated on a most-favored basis. In any case, the key issue is whether the level of risk has in fact increased or not. We are currently discussing this point with the U.S., but at this stage, we have not yet received a definitive assessment as to whether the risk has indeed increased. However, from our perspective, we need to take the worst-case scenario into account, and in that sense, we cannot rule out the possibility that the risks have increased. Therefore, we believe it is necessary to proceed with our efforts without delay. That said, looking at Company A and Company B, Company A has not yet been able to fully implement technical safeguards, including guardrails, and as a result, it has temporarily suspended general availability. On the other hand, Company B has been conducting testing with us within a limited-release framework. Given this, what we can realistically do as the Japanese authorities is to ask whether it is possible to significantly accelerate this testing process and move more quickly toward practical and secure implementation. That is what we believe is the best course of action available at this time.