- Japan is weighing penalties for AI companies that refuse to cooperate with the government.
- The target is largely the US frontier labs, such as Anthropic and OpenAI, that Japan depends on but cannot currently force to share information.
- The open question is whether the fines will be big enough to matter or as small as South Korea’s roughly $21,000 cap.
Japan is considering penalties for AI companies that ignore government orders.
The penalties will be aimed mostly at U.S. developers of advanced AI, whose systems Japan uses but cannot force to share information. Officials plan to publish a full reform blueprint before the end of 2026.
Why can’t Japan make American AI companies cooperate?
Japan’s AI Promotion Act, which states that if AI misuse harms someone’s rights, the government can investigate and give guidance, took full effect in September 2025. But the law has no punishment for a company that simply refuses to cooperate. Japan is now considering including penalties in its law.
This regulatory gap is a bigger problem with foreign companies. A group within Japan’s ruling Liberal Democratic Party, led by former digital minister Masaaki Taira, warned in a draft proposal that other governments can get information from AI companies, but Japan may not be able to. The draft noted that the European Union already fines AI companies that refuse to hand over information.
Japan is most worried about advanced AI that it does not build or control. The government wants AI companies to take on safety duties. It is also thinking about penalties for the risk of advanced AI going out of control or being used in cyberattacks.
A private working group has confirmed this plan. Officials also want the power to act against companies that refuse to help after something goes wrong.
Japan’s Financial Services Agency reportedly set up a 36-member group of government and industry members in May 2026 to study cyber threats linked to Anthropic’s Mythos model. The Japanese branches of Anthropic and OpenAI are taking part.
Notably, Japan’s updated AI Basic Plan, approved in July 2026, named high-performance models such as Claude Mythos as a concern. The U.S. Commerce Department also limited access to Anthropic’s Mythos 5 and Fable 5 earlier in 2026, then restored it.
The government reportedly began reviewing the AI Promotion Act at an expert panel meeting on October 26. Digital minister Toshiharu Furukawa told the panel that Japan needs closer work with partner countries on AI safety. A group of academics held its first meeting two days later, and the panel hopes to settle on a direction by the end of the year.
It is not yet clear that penalties will survive this process. Some participants have suggested that the country proceed with caution, because other countries are not all considering tighter AI rules.
Will Japan’s penalties be big enough to matter?
Earlier this year in January, South Korea’s AI Basic Act that lets regulators fine companies up to 30 million won (about $21,000) for refusing inspections or ignoring corrective orders took effect. However, for a large U.S. AI company, a fine that small is negligible and unlikely to change behavior.
On the other hand, as part of an effort to make Japan the “most AI-friendly country,” its first AI law, the AI Promotion Act that was passed in May 2025, left out penalties so that cautious firms would not be scared away from using AI.
The law was meant to work through social pressure and voluntary compliance, and analysts assumed that if a company like OpenAI wanted access to Japan’s market, it would have to follow local rules.
By the end of the year, it will be clear how large Japan’s fines will be, or if there will be any at all.
If you're reading this, you’re already ahead. Stay there with our newsletter.
FAQs
Why does Japan want to add penalties to its AI law?
Because the current AI Promotion Act lets the government investigate and advise but cannot punish companies that refuse to cooperate, leaving it unable to compel information from operators, especially foreign ones, the way the EU can.
Which companies would the penalties target?
The rules would focus on frontier AI developers behind high-performance systems, mostly US labs such as Anthropic and OpenAI, whose models like Claude Mythos Japan uses but does not control.
How do Japan's planned penalties compare with South Korea's?
South Korea's AI Basic Act, effective January 22, 2026, caps fines at about US$21,000 (30 million won) for refusing inspections or ignoring orders; Japan has not set its own figure and plans to outline its approach by the end of 2026.
Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.
Hannah Collymore
Hannah is a writer and editor with nearly a decade of blog writing and event reporting experience in the crypto space. At Cryptopolitan, Hannah contributes to the news page, reporting and analyzing the latest developments in DeFi, RWA, crypto regulation, AI and frontier tech industries. She graduated from Arcadia university with a degree in Business Administration.
