[CLSTR]
< Back to situation

[REVISION HISTORY]

Japan AI regulation and personality rights reforms

Updated 4 times since CLSTR started tracking revisions of this situation.

What changed

2026-08-20 07:29 UTC → 2026-08-27 01:31 UTC · added removed

In July 2026, the Japanese Diet enacted several AI-related reforms. An amendment to the Personal Information Protection Act now permits sharing sensitive data—such as medical history and criminal records—for AI training or statistical purposes without individual consent. While supporters claim this will accelerate innovation, opposition parties warned of data leak risks and potential threats to Japan's EU-GDPR adequacy status. Lawmakers also passed measures to combat election misinformation. Amendments to the Public Offices Election Law and the Information Distribution Platform Law require AI-generated political content to be clearly labeled. Social-media platforms are now obligated to curb false information and disclose AI-generated images or videos of candidates. These rules, backed by the LDP and Japan Innovation Party, take effect on 1 March 2027, ahead of unified local elections. Addressing the rights of creators, the Justice Ministry's expert panel proposed civil liability for the unauthorized use of famous individuals' voices and likenesses. This follows a campaign by voice actress Megumi Ogata, who estimated losses of ¥2–4.5 billion due to unauthorized AI voice models. In early August 2026, the Ministry issued guidelines treating a person's voice with the same commercial protections as their name or image. Furthermore, a To protect intellectual property, the government panel has approved adopted a draft non-binding ‘principle code’ to protect intellectual property rights. for generative AI businesses, including foreign entities. Utilizing a ‘comply or explain’ approach, the nonbinding framework urges generative AI businesses—including foreign entities operating in Japan—to disclose training data and collection methods. The code includes requirements for businesses encourages operators to make disclose training data accessible on websites data, collection methods, and to disclose specific webpage usage to rights holders alleging infringement. Intellectual property strategy minister Kimi Onoda stated the government will monitor the situation and consider further measures if necessary. In the private sector, creative services company Somethingfun Co., Ltd. implemented its own 13-item AI policy on 30 July 2026, which includes preventing deepfakes and offering ‘Non-AI’ production options.

Versions

  1. 2026-08-27 01:31 UTC Japan AI regulation and personality rights reforms
  2. 2026-08-20 07:29 UTC Japan AI regulation and personality rights reforms
  3. 2026-08-09 03:15 UTC Japan AI regulation and personality rights reforms
  4. 2026-07-30 01:38 UTC Japan AI regulation reforms
  5. 2026-07-29 08:27 UTC Japan AI regulation reforms

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.