Japanese banks get OpenAI GPT-5.5 for cybersecurity use

Several Japanese financial institutions have been granted access to OpenAI’s GPT-5.5 model to bolster their cyber defenses, Finance Minister Satsuki Katayama said after meeting OpenAI’s chief strategy officer in Tokyo.

The technology is intended to help banks improve detection, monitoring and response to more advanced cyber threats. Katayama described the arrangement as ‘a welcome development’ and said it would assist institutions in defending against cyberattacks.

A separate report named MUFG Bank, Sumitomo Mitsui Banking Corp and Mizuho Bank as expected partners. The same report said GPT-5.5 is available only to trusted partners and indicated the model is comparable in capability to a leading alternative developed by another company. Neither the finance ministry nor OpenAI confirmed the list of participating institutions during the minister’s remarks.

OpenAI operates through a hybrid structure that includes a nonprofit foundation and a commercial entity. The company has a long-standing technical partnership with Microsoft that has supported broader distribution of its models. Japanese officials and industry representatives have been discussing how advanced models can be used safely within critical systems such as banking infrastructure.

Officials have not disclosed detailed plans for how the model will be deployed, trained or overseen inside financial firms. Timelines for testing, integration and regulatory review have not been released.

Banks and regulators worldwide have increased investment in tools to detect and respond to cyber incidents as attacks become more complex. In Japan, authorities view access to an advanced generative AI model as one option to strengthen resilience across the financial system.

The announcement follows talks between Japanese officials and OpenAI representatives about applying advanced AI within the country’s financial sector. Operational details and the full list of participating firms remain undisclosed.