Japan’s government is to seek early information on undisclosed software vulnerabilities from US AI vendors and other overseas technology companies as part of a broader response to a spike in cyber-attacks this year, Nikkei reported.

An emergency cyber-security package is to be presented at a ministerial cyber-security meeting planned for as early as Friday, after prime minister Sanae Takaichi instructed officials to develop measures at a meeting in late October, the report said.

The plans follow a rise in cyber-attacks in Japan this year that take advantage of new open-source and AI-powered hacking tools, some of which are believed to be carried out by Russian groups that may at times be working with North Korean threat actors, according to TrendAI, a unit of Tokyo-based Trend Micro.

A recently disclosed vulnerability in PHP also led to widespread patching of online systems by Japanese companies.

Japan’s government and companies already have access to advanced AI models from US companies such as Anthropic and OpenAI, and the government recently asked for broader access to those models to help respond to cyber-attacks, Nikkei’s report said.

The government also reportedly plans to build a broad information-sharing framework, collecting intelligence from companies that have been attacked as well as cyber-security vendors, cloud service providers, and other organisations involved in cyber-security in order to encourage the disclosure of data.

Public-private coordination

The information is reportedly to be centralised at Japan’s National Cybersecurity Office, enabling authorities to grasp the overall scope of attacks across industries and regions and assess risks.

The package is also reportedly to include measures to encourage cyber-security vendors and cloud service providers to join public-private coordination councils.