SoftBank Group (SBG) announced at a corporate client event in Tokyo on June 16 that it will launch a service using AI technology to combat cyberattacks, targeting social infrastructure industries such as finance and information and communications in Japan. The service will employ cutting-edge models from OpenAI. It will first assess corporate network security vulnerabilities and, in the future, plans to offer a function for writing remediation programs. Chairman and President Masayoshi Son stated at the event: “AI-powered cyberattacks pose a crisis for Japan, the likes of which we haven’t seen since the Black Ships.” SoftBank Group did not disclose the specific model version used for the service; in a media interview before the event, Son said: “OpenAI has already built trust relationships with governments and businesses.”
The announcement comes against a backdrop in which Anthropic’s high-performance AI model, Claude Mythos, was recently forced to suspend external access due to U.S. government export control directives. Concerns are growing that if similar capabilities were misused, they could throw financial institutions and others into chaos, and parties are moving urgently to explore countermeasures. By choosing OpenAI’s model as its foundation, SoftBank Group is also, to some extent, avoiding the aforementioned uncertainty. The service has already begun to be gradually introduced to target industry clients, and applications are being accepted.