Simultaneous cyberattacks are spreading across Japan, affecting a wide range of sectors, including universities, restaurants and media. The spread of open AI models and a state-backed attack infrastructure using AI have been cited as possible factors.
Simultaneous cyberattacks are occurring across Japan at an unprecedented pace. At 物語コーポレーション, which operates Yakiniku King, unauthorized access to its official app has raised the possibility that member information was compromised. At Osaka Metropolitan University, approximately 500 servers were taken offline, and personal information on as many as approximately 130,000 people may have been leaked.
The affected sectors range from universities, restaurants, media, manufacturing, financial services and logistics, and the scope is expanding. The cause and details remain under investigation, but the spread of open AI models and the possibility of automated hacking using AI have been cited as potential factors.
The rapid spread of frontier-level open-weight AI models from China, including DeepSeek, GLM and Kimi, is said to be a factor, with concerns that they could be readily misused because they can be operated in local environments and used to evade monitoring. Techniques for illicitly distilling high-performance AI models have also been identified as a concern, raising the possibility that advanced capabilities could be transferred to open models through unauthorized training.