OpenAI to provide Japan cybersecurity model

- OpenAI said on May 21 it will provide Japan’s government and some companies with a cyber-defense model as concern grows over AI-enabled attacks. - Paul Nakasone, the former NSA chief now on OpenAI’s board, joined the Tokyo push as OpenAI promoted a model called GPT-5.5-Cyber. - OpenAI said it hopes Japan’s government will be an early partner as talks continue with officials and selected companies.

OpenAI said on May 21 that it will provide the Japanese government and some Japanese companies with an artificial intelligence model built for cyber defense, adding a national-security customer track to a company better known for ChatGPT. The announcement came in Tokyo as governments and companies weigh how generative AI can both enable and help stop cyberattacks. Japanese reports said the model would be offered to help defend critical systems and other high-priority networks. OpenAI has separately been promoting new research claims and is also the subject of reports that it may confidentially file for an initial public offering as soon as Friday. ### Which model is OpenAI offering Japan? The Yomiuri Shimbun’s Japan News and Asahi Shimbun’s English-language service reported that OpenAI plans to provide a cyber-focused model described as “GPT-5.5-Cyber” to the Japanese government and some companies. The reports said the system is designed for defensive cybersecurity work rather than general consumer use. Japan Today, citing Kyodo, reported that OpenAI said Thursday it would provide an AI model with advanced cybersecurity capabilities to the Japanese government and some companies. The move comes as concern grows over cyberattacks that use AI, the report said. ### Who is making the pitch in Tokyo? Paul Nakasone, the retired U.S. general who led the National Security Agency and U.S. (japannews.yomiuri.co.jp) Cyber Command, was part of OpenAI’s outreach in Japan, according to reporting surfaced on May 22. Sasha Baker, OpenAI’s head of national security policy and a former senior Pentagon official, was also involved in talks with Japanese counterparts, according to that reporting. (article.wn.com) OpenAI’s own cyber program gives context for the Japan offer. On April 14, the company said it was expanding its Trusted Access for Cyber program, introducing GPT-5.4-Cyber to vetted defenders and hundreds of teams responsible for protecting critical software. OpenAI said the program was built around verified defenders and added safeguards for high-risk cyber capabilities. (kantenna.com) ### Why is Japan in the frame now? Japan’s National Center of Incident Readiness and Strategy for Cybersecurity says the government’s strategy calls for deeper practical cooperation with allies and like-minded countries on cyber defense. That policy backdrop has made cybersecurity a standing issue for Japanese ministries and companies that run critical infrastructure. (openai.com) Asahi reported that OpenAI plans to provide the model soon to some Japanese companies, and said the company described its performance as comparable to Anthropic’s Claude Mythos in some respects. The report said the focus was on strengthening defenses as AI tools become more capable in offensive cyber operations. ### What else is OpenAI saying about its technical progress? (cyber.go.jp) OpenAI said on May 20 that one of its models had solved the planar unit distance problem, a question first posed by Paul Erdős in 1946. In a post on its website, the company said the model produced a result that disproved a major conjecture in discrete geometry and that mathematicians had checked the proof. (asahi.com) The claim added to OpenAI’s broader argument that its reasoning systems are improving beyond consumer chat tasks. OpenAI described the result as a research milestone, while outside reports said the work involved a general-purpose reasoning model rather than a system trained only for mathematics. ### Where does the IPO report fit in? (openai.com) CNBC reported on May 20 that OpenAI was preparing to confidentially file a draft IPO prospectus as soon as Friday. CNBC said the company could be headed for one of the largest public listings in history and reported that an OpenAI representative said the company’s focus “remains on execution.” Other outlets, including CNBC TV18, reported that private investors valued OpenAI at more than $850 billion and said Goldman Sachs and Morgan Stanley were working on a potential listing. (openai.com) Those details have not been confirmed in an OpenAI filing. OpenAI has not yet publicly posted any IPO registration documents. (cnbc.com) The next concrete milestone would be a confidential filing becoming public later in the process, while the Japan cyber rollout is expected to proceed through talks with Japanese government agencies and selected companies. (cnbctv18.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.