Several Japanese financial groups have been granted access to OpenAI’s GPT-5.5 model for use in cyber defence, Finance Minister Satsuki Katayama said after talks in Tokyo with the company’s chief strategy officer, reported Reuters.
The news agency noted that more advanced AI systems with strong coding ability have increased concern among policymakers and companies because they can also help attackers find software weaknesses and plan intrusions.
A possible countermeasure for banks and other organisations facing disruption is to receive early use of such models under restricted access arrangements for selected partners.
Katayama did not identify the institutions involved, but said the move was “a big step forward in strengthening Japanese financial institutions’ ability to defend against cyberattacks”.
She made the remarks to reporters after meeting Jason Kwon, OpenAI’s chief strategy officer.
Recently, Nikkei reported that the country’s three largest banks, MUFG Bank, Sumitomo Mitsui Banking Corp and Mizuho Bank, were likely to receive OpenAI’s newest model, which is seen as comparable to Anthropic’s Claude Mythos.
Katayama also said Japan’s government and financial institutions were expected to receive access to Mythos to improve their cyber defences.
Earlier this month, Japan created a public-private working group to examine cybersecurity threats to the financial system linked to Mythos, as authorities step up efforts to respond to AI-related risks.
The decision by OpenAI to extend access to Japanese financial institutions came after discussions between Tokyo and Washington, following similar access already provided to European companies.
Earlier this month, the Financial Times reported that Anthropic agreed to brief major finance ministries and central banks on cybersecurity weaknesses in the global financial system identified by its latest AI model, citing unnamed sources.
