July 16, 2026
TOKYO – The government approved the revisions to its Artificial Intelligence Basic Plan on Tuesday, laying out a policy to greatly strengthen its ability to counter cyberattacks in light of the emergence of high-performance AI systems such as Claude Mythos.
The revised plan calls for ongoing reviews of the legal framework with an eye on establishing penalties for businesses that infringe on the rights of citizens.
The initial Basic Plan, which serves as a guideline for AI policy, was approved by the Cabinet in December, and the first revision was approved a little over six months later.
The revised plan points out that “the technological and security risks posed by AI are becoming more complex and severe,” particularly the risks posed by cyberattacks due to the misuse of AI. The plan states that Japan will “continuously strengthen its cyber countermeasure capability.”
The government aims to work more closely with foreign government agencies and AI developers to gain early access to the latest models, which are often not widely released. It will actively utilize high-performance AI to assess vulnerabilities of government systems and expand the capabilities of the Japan AI Safety Institute, a government-affiliated agency responsible for evaluating AI performance.
The current AI law stipulates that if citizens’ rights are infringed upon due to the misuse of AI, the government will investigate and provide guidance and advice to the operators. However, no penalties are stipulated for operators who refuse to cooperate with such investigations.
In the revision, the government added the clause that it “will proactively and continuously review the system, including the AI law, to deal with risks more effectively,” thereby opening up the possibility of strengthening regulations, such as the introduction of penalties.