Japan's revised Act on the Protection of Personal Information passed the upper house plenary on July 10, 2026. It creates a "special provision" allowing companies to share personal data without the individual's consent when the purpose is statistics creation or AI development, while at the same time introducing the country's first surcharge (administrative fine) system, combining deregulation and tightening in a single package.
July 10, 2026 · Act on the Protection of Personal Information (Revised)
Japan Clears AI Data-Sharing Without Consent — and Adds Its First Privacy Fines
The upper house passed a bill that lets companies share personal data without individual consent for statistics creation and AI development, while introducing the country's first surcharge system — deregulation and tightening bundled into a single package.
1st
surcharge (administrative fine) system for privacy violations in Japan
<16
new protections codified for minors under 16, covering legal guardians
~2028
expected enforcement — within 2 years of promulgation
Path Through the Diet · 2026
Apr 7
Cabinet approves & submits to the Diet
→
May 26
Passes lower house plenary
→
Jul 8
Clears upper house special committee
→
Jul 10
Enacted in upper house plenary
What's In the Package — One Carve-out, Four Tighteners
Deregulation
Statistics-creation special provision
Enables no-consent third-party data provision for statistics purposes, including AI development.
Surcharge system
First-ever administrative fines on violating businesses.
Protection for under-16s
Codifies rules covering legal guardians.
Specific biometric data
Notification duties for facial-feature data.
Breach notification
Eased where harm-to-rights risk is low.
Proponents
Smoother AI development and statistics creation should ease data collection for companies and research bodies — powering new services and progress in medicine and drug discovery.
Critics
Sensitive data — medical history, beliefs, criminal records — could flow to AI development without consent. Concerns: weak sanctions, hard-to-check purpose limits, no collective litigation system.
The consent-first principle had been flagged as a practical barrier to AI development such as LLM training.
A supplementary resolution urged measures to prevent identification of individuals and thorough oversight of contractors — with practitioners pushing to strengthen governance before enforcement begins.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…