Japan introduces AI transparency principles to protect intellectual property

The Japanese government has adopted guiding principles to protect intellectual property rights in the field of generative artificial intelligence. The non-binding Principles Code encourages AI operators, including those based overseas that provide services in Japan, to disclose the types of data and methods used to train their models. This initiative aims to increase transparency and address concerns regarding the unauthorized use of copyrighted texts and images.

In response to the evolving technological landscape, Kimi Onoda, the minister for intellectual property strategy, stated that the government will monitor the situation and consider further measures if necessary. Companies that adopt the code will publicly outline their learning processes and data collection methods on their websites.

Complementing these regulatory efforts, the creative services company Somethingfun Co., Ltd. has established its own generative AI policy. Effective July 30, 2026, the policy includes 13 items focused on protecting the rights of creators and performers, preventing deepfakes, and ensuring transparency. The company also explicitly offers ‘Non-AI’ production options for clients who prefer not to use generative tools.