Japan's Digital Agency built GENAI, a sovereign generative-AI platform deployed to all Agency staff in May 2025. One documented case cut a 2-month task to 3 days. Codebase released as OSS; large-scale pilot covering 180,000 civil servants across all ministries launched January 20
2 months to 3 days
Task completion time reduction (case study)
~180,000
Civil servants in FY2026 pilot (Release 1.0, from January 2026)
Details
Maturity
Scaling
Promoter
Digital Agency (Dejitaru Cho), Government of Japan
Period
2025-present
Keywords
generative AI, large language models, government productivity, open-source, public administration
Context
GENAI (Government AI) is a sovereign generative-AI platform developed in-house by Japan's Digital Agency and deployed to all Agency employees in May 2025.
Objectives
To integrate large language models into domain-specific government applications, including the 'Diet Response Search AI' for retrieving relevant parliamentary Q&A records and the 'Legal Research Support AI' for accelerating drafting and checking of administrative legal documents.
Activities
Further deployed uses include AI-based classification of large volumes of public comments (from June 2025) and AI-driven analysis of administrative documents across ministries (from February 2026). The GENAI codebase was released as free open-source software under commercial-use-permitted licences. Governance runs through a published 'Guideline for Japanese Government Procurement and Utilization of Generative AI', an inter-ministerial steering process, mandatory security review, and a monitoring framework for privacy, output quality and labour-saving.
Results
The Digital Agency's official report documented a case where a task previously requiring two months was completed in three days with AI assistance. In fiscal year 2026, a large-scale pilot extended GENAI to approximately 180,000 civil servants across all Japanese ministries and agencies (Release 1.0, from January 2026), with full national implementation targeted for FY2027.
Conclusions
The platform keeps government data within national infrastructure and avoids single-provider dependency.
Implementation
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Data sources
Where this practice's information was retrieved from, and when.