The Japanese government is pressing ahead with a legal revision that would allow sensitive personal data such as medical records and criminal histories to be used for artificial intelligence (AI) development without the person’s consent.

Cryptopolitan, a blockchain media outlet, reported on June 6 that Japan Digital Minister Hisashi Matsumoto (히사시 마쓰모토) said Japan could become an "AI colony" if it fails to narrow the gap with global competitors.

At a news conference the same day, Matsumoto said Japan must speed up AI development and stressed the need for the bill. "I hope many Japanese people will understand that we must keep pushing AI development," he said. "Otherwise, Japan could become an 'AI colony'." The revision has already passed the lower house and now awaits deliberation in the upper house.

The key issue is the scope of access to data for AI training. The revision would open the way for sensitive personal information, including medical records and criminal histories, to be used in AI development for statistical purposes. Opposition parties are pushing back, saying the measure could weaken personal data safeguards and increase the risk of data leaks. Matsumoto said the revision does not lead to personal data leaks and that expanded data access would be limited to statistical use cases related to AI development.

Japan’s move to adjust regulation comes amid an investment gap with the United States and China. From 2019 to 2023, U.S. domestic AI research investment totalled about $329 billion, while China invested about $133 billion, but Japan was at about $10 billion. The Japanese government has changed subsidy rules and provided funding to narrow the gap, and has also pursued related legal revisions.

It is also pursuing cooperation with foreign companies. Microsoft and OpenAI expanded cooperation with Japan under the framework of the U.S.-Japan security alliance. The Japanese government is also supporting SoftBank, Sakura Internet and Japanese semiconductor companies to build domestic AI models and computing infrastructure. OpenAI visited Japan in late May and proposed a cybersecurity-focused AI system, "GPT-5.5 Cyber," to the government and private companies. Paul Nakasone, an OpenAI board member and former head of U.S. Cyber Command, said he discussed defensive measures in 15 key areas with Japanese authorities.

Still, scepticism remains within Japan about building a fully domestic AI system. When the Ministry of Economy, Trade and Industry proposed using public funds to develop a Japan version of ChatGPT, some in the ruling party responded that it was reckless because Japan lacked the resources to compete with U.S. and Chinese companies. The ministry ultimately dropped the goal.

Instead, the Japanese government plans to revise its basic AI plan this summer. A draft has raised the possibility of further strengthening AI sovereignty provisions tied to national security. Within the ruling Liberal Democratic Party’s Digital Society Promotion Headquarters, the argument is also gaining traction that Japan should diversify suppliers rather than build an entire domestic system from scratch. Secretary-General Akihisa Shiozaki (시오자키 아키히사) said last month, "What matters most is ensuring autonomy without relying on any one country, company or provider."

The debate is not only a Japanese issue. The European Union has also recently rolled out a technology sovereignty package to strengthen its domestic cloud, AI and semiconductor industries to reduce reliance on U.S. technology companies. Japan’s debate over the bill is also serving as a test bed for how far it will change rules between securing AI competitiveness and protecting personal data.