AI Basic Plan Revised in a Short Term

Sanae Takaichi Cabinet revised Japan’s AI Basic Plan to enhance risk management of artificial intelligence. While it proposes strengthening systems of the government, including formulating a guideline for using AI or regulating legislation, the revised plan does not offer any specific measure on how to deal with cyber attacks. It is not easy for the government of Japan to establish “AI sovereignty” against foreign countries. 

The government firstly issued AI Basic Plan in December 2025. It was overwhelmingly focused on how to promote AI technology in Japan. Takaichi promised to “make Japan the world’s best country to develop and use AI” in her policy speech to the Diet in October 2025. She simply believed that promotion of AI would accelerate economic growth through active investment on it. AI was mainly considered as one of the tools for implementing her economic policies.

 

Development of AI Mythos by Anthropic significantly changed the situation of Takaichi government’s AI policy. AI Mythos has an advanced capability in discovering vulnerability in software or system, threatening security of system owned by Japanese government and crucial industries in Japan. Although Japan obtained access to AI Mythos to find and fix vulnerability in systems, the government of United States prohibited foreign countries using AI Mythos in the context of export control. The government of Japan faced a necessity to improve security of its own AI.

 

Takaichi government decided to revise AI Basic Plan published only six months ago. New plan expects establishment of AI sovereignty with independence on specific country or corporation. To achieve credible AI, the government hopes to lead enhanced management of data or development of fundamental models. It believes that Japan can develop its own AI that can solve not only problems in Japan but in the world.

 

To achieve “responsible agile governance,” the government is going to frequently revise existing legal systems including AI Technology Promotion Act. The law gives the government an authority to take measures against damages on public interest brought by inappropriate development of AI. The government will make new rules for introducing AI agents to Japan. The guideline will coordinate conflict of interests related to AI.

 

The government established AI Safety Institute (AISI) in 2024. The organization assesses safety of AI, cooperating with ministries and foreign institutes. The revised AI Basic Plan welcomes further effort of AISI to improve capability for collecting international information on discovering vulnerability in systems. The plan also expects Project YATA-Shieldto check out vulnerability in system owned by the government.

 

However, the plan shows limit of bureaucratic measures on cutting-edge technology. Bureaucrats are well at creating or managing framework, including conferences or legislature. But all they could do to the threat of AI Mythos was not creating Japanese version of AI Mythos, but asking Anthropic Japan’s access to AI Mythos. Takaichi government likes something “number one” in the world. It is therefore necessary for them to have broader coordination for AI security to make Japan the world’s best country to develop and use AI.

Comments

Popular posts from this blog

LDP Kickoffs Presidential Election

Defense Ministers Meet in Seoul

Contaminated Soil to Prime Minister’s Residence