Japan has national AI legislation. That fact alone sets it apart from the United States, which still lacks a federal AI law. But the more consequential distinction is how Japan’s framework is designed: not as a restriction, but as a promotion.
The Act on Promotion of Research and Development and Utilization of Artificial Intelligence-Related Technologies, Japan’s AI Act, reportedly became effective in May 2025, according to legal analysis of the legislation. A Basic AI Plan was reportedly approved in late 2025 to operationalize the Act’s principles, according to international legal trackers. The Act itself does not impose prescriptive duties on AI developers or deployers. There are no mandatory risk assessments, no prohibited AI categories, and no enforcement penalties tied to non-compliance with specific technical requirements.
Multiple independent legal analysts characterize the Act as a “soft law” framework: principles-based, voluntary in application, and structured around guidance rather than mandate. Bird & Bird’s analysis describes it as a soft-law instrument; the International Bar Association characterizes it similarly. Japan’s government has signaled an ambition to position the country as a leading destination for AI development, a policy aspiration, not a statutory requirement, but one that shaped the law’s design.
This matters for compliance professionals because the contrast with the EU AI Act is structural, not just philosophical. The EU’s framework is precautionary and rights-based: it restricts certain uses, mandates documentation, and creates enforceable obligations. Japan’s framework is promotion-first: it encourages AI development and expects voluntary adherence to stated principles. An organization operating in both jurisdictions faces two entirely different compliance architectures. In the EU, the question is “what does the law require us to do?” In Japan, the current question is closer to “what does the government hope we’ll do?”
The practical implication is not that Japan requires nothing, it’s that Japan’s requirements, for now, sit in a different register. The Basic AI Plan provides operational guidance under the Act. Companies with Japan operations should track guidance issued under that Plan, even if no binding obligation currently attaches to non-compliance.
What to watch: Japan’s soft-law approach may shift as AI applications scale. The government has framed the current framework as a starting point, not a ceiling. Guidance issued under the Basic AI Plan is the layer most likely to move first. Compliance teams should monitor for any transition from voluntary guidance to prescriptive rule, a step several jurisdictions have taken after initial soft-law phases.
The TJS takeaway: Japan’s AI Act is not a compliance burden in the conventional sense. It is, however, an important data point in the global regulatory map. As organizations build multi-jurisdictional AI governance postures, understanding Japan’s promotion-first model clarifies why there is no one-size-fits-all compliance framework. The EU demands conformity. Japan requests alignment. Those are different obligations, and they require different internal responses.