Japan's cabinet approved the country's first-ever AI basic plan on December 19, with Prime Minister Sanae Takaichi declaring it time to "launch a counteroffensive" through public-private cooperation. The plan targets raising generative AI adoption from the current 26.7% to 80% of the population, while attracting roughly ¥1 trillion ($6.4 billion) in private sector investment.
The adoption problem
The numbers are stark. A government white paper released in July found that only 26.7% of Japanese individuals used generative AI in fiscal 2024. China sits at 81.2%. The US at 68.8%. That gap is massive, and it's not simply demographics, though Japan's aging population doesn't help.
Even among people in their 20s, Japanese adoption reaches only 44.7%, which sounds reasonable until you realize this is the group most likely to use new technology. The white paper also found that among companies, just 49.7% of Japanese firms planned to use generative AI, compared to over 80% in China and the US.
Why so low? The survey found that more than 40% of respondents admitted they "don't know how to use it," and nearly 40% said generative AI "is not necessary in my everyday life."
What the plan actually says
The approved plan positions AI as "social infrastructure," which is bureaucratic language for something the government considers essential. The draft, which was finalized before the cabinet meeting, describes AI as "an intellectual foundation and an execution platform."
Beyond the adoption targets, the plan calls for creating an "AI ecosystem" where developers, users, semiconductor manufacturers, and cloud providers collaborate. The government wants to use this ecosystem to expand into overseas markets and address Japan's digital trade deficit.
Every ministry and agency will be expected to utilize AI, with the goal of eventually reaching all government employees. The ¥1 trillion private investment target includes funding for human resources, infrastructure development, and supporting AI industry expansion into emerging markets.
Takaichi's framing matters here. In her October policy speech, she pledged to make Japan "the easiest country in the world to develop and utilize AI." The December 19 announcement emphasized "reliability as a core Japanese value," which sounds like positioning for future international AI governance discussions. She also instructed ministers to work toward hosting an "AI Summit" in Japan.
The bigger investment picture
This plan fits into a much larger funding story that's been building throughout 2025. The government announced an "AI and Semiconductor Industry Strengthening Framework" in late 2024 planning for ¥10 trillion ($65 billion) in public support by 2030. METI's fiscal 2025 budget includes ¥330 billion for AI and semiconductors.
Some context on that ¥10 trillion figure: a significant portion goes to next-generation semiconductor development, not AI specifically. The government has provided ¥920 billion to Rapidus for its Hokkaido semiconductor factory. METI awarded ¥72.5 billion to five companies for AI supercomputer development, with Sakura Internet receiving the largest allocation at ¥50.1 billion.
Private investment tells a different story. According to Stanford's 2025 AI Index Report, Japan's private AI investment in 2024 totaled just $0.93 billion. The US invested $109.1 billion. Even the UK managed $4.5 billion. Japan ranked 12th globally.
The regulatory angle
Japan's approach to AI governance deliberately avoids EU-style hard regulations. The AI Promotion Act, which Parliament approved in May 2025 and took effect in June, contains no penalties for noncompliance. The only obligation on businesses is to "endeavor to cooperate" with government measures.
This is intentional. The government wants Japan to be, in its own words, the "most AI-friendly country" for development. The AI Guidelines for Businesses, updated in March 2025, remain non-binding soft law.
Critics argue this approach lacks specificity. The AI Safety Institute, which is supposed to provide oversight, faces questions about whether it has sufficient authority or resources to address misuse. Annual updates are planned to keep pace with the technology, though whether guidelines can move fast enough remains to be seen.
