Particle.news

Japan Adopts AI Transparency Principles

The non‑binding rules ask generative‑AI firms to disclose how they collect training data and aim to curb copyright misuse while leaving enforcement to future policy or industry action.

Overview

  • The government formally adopted basic principles on Tuesday that require generative‑AI companies, including overseas firms such as Google, to disclose summaries of how training data are collected to reduce copyright infringement.
  • A three‑member cross‑party Diet delegation traveled to China to seek meetings with the Communist Party’s external‑relations officials, a limited reopening of parliamentary contact that was halted after tensions late last year.
  • The National Police Agency published its 2026 white paper and for the first time acknowledged multiple investigative failures in the Ōkawahara wrongful‑conviction case and outlined steps to prevent similar errors.
  • Local political accountability continued to move forward as Morioka city councilor Tomoki Yamazaki was sent to prosecutors over alleged improper use of political activity funds and apologized while refusing to resign, and Fukuoka assembly chairman Isuo Kurauchi announced his resignation amid money‑transfer allegations.
  • Regional and safety updates included Miyagi Prefecture reporting record 2025 tax revenues tied to the end of a flat tax cut and higher consumption tax receipts, and the environment and consumer agencies issuing recommended performance standards for bear‑repellent spray to improve public safety.