The EU AI Act's transparency obligations will begin applying in August 2026. The differences from Japan's AI Promotion Act and key points of corporate governance are explained, including the risks of extraterritorial application, the treatment of high-risk AI and potential future legislative amendments.
As business use of AI, including generative AI, expands, the EU will begin applying the AI Act's transparency obligations in August 2026. This could affect Japanese companies that provide AI-enabled products and services in the EU market. This article outlines the differences between Japan's AI Promotion Act and the EU AI Act.
EU law entails legal obligations and enforcement measures, and violations may result in fines of up to €35 million or 7% of worldwide annual revenue. By contrast, Japan's AI Promotion Act emphasizes innovation by prioritizing research and development and the promotion of use, and does not provide for sanctions such as fines. This is a major difference.
With regard to the risk of extraterritorial application, Japanese companies that sell or provide AI-enabled products in the EU may incur obligations depending on the business activities and applications involved. Applications in fields such as automotive, robotics and machinery have been identified as potentially involving high-risk AI. Understanding which AI is used for which business activities and how it is used is considered more important than whether the AI was developed in-house.
The article also organizes key points for companies to consider regarding AI governance, taking into account insights from NAVEX's domestic survey on identifying risks and strengthening governance in Japan, as well as future legislative developments such as the amended Whistleblower Protection Act, which takes effect on 1 December 2026.