Japan Issues New AI Transparency Code for Training Data

The Japanese government adopted a new non-binding code of conduct for generative AI operators on August 25, 2026 ^[1][2]. The guidelines urge companies to disclose outlines of their training data and collection methods ^[1]. This initiative aims to improve industry transparency and protect intellectual property rights ^[1][3].

The code specifically targets operators of large-scale models ^[2]. While the rules are currently voluntary, officials hope they will foster trust among creators and users ^[1]. The move follows global discussions on the ethics and legalities of AI data sourcing ^[3].