TOKYO (Reuters) - OpenAI has given some Japanese financial institutions access to its GPT-5.5 model to help prevent cyberattacks, Japanese Finance Minister Satsuki Katayama said on Friday after a meeting with the U.S. company's chief strategy officer.
The release of new models of AI able to code at a high level has given hackers an unprecedented ability to identify cybersecurity risks and devise ways to exploit them.
One defence for the banks and other companies at risk of being destablilised is early access to the new models that are available only to trusted partners.
Without naming any of the financial institutions, Katayama said the access they were getting was "a big step forward in strengthening Japanese financial institutions' ability to defend against cyberattacks". She was speaking to reporters after meeting Jason Kwon, OpenAI's chief strategy officer, in Tokyo.
The Nikkei newspaper reported on Thursday that Japan's three biggest banks - MUFG Bank, Sumitomo Mitsui Banking Corp and Mizuho Bank - were expected to gain access to OpenAI's latest model, which is believed to be on a par with that of rival Anthropic's Claude Mythos.
MUFG and Mizuho declined to comment. Sumitomo Mitsui Banking was not immediately available to comment.
Katayama said on Friday that Japan's government and financial institutions were also expected to gain access to Mythos to enhance their defences.
As it strives to counter threats from AI, Japan this month set up a public-private working group to address cybersecurity risks to the financial system posed by Mythos. OpenAI's decision to grant access to Japanese financial institutions followed negotiations between Japanese and U.S. governments after it allowed European companies similar access.