Qlean Dataset Strengthens Data Support System for Digital Agency's "Government AI" Selected Companies, Bolstering Trust in Administrative DX
Visual Bank Inc.'s "Qlean Dataset" has strengthened its data support system specifically for administrative practices, targeting companies selected for the Digital Agency's "Government AI (Gennai)" project. This initiative aims to powerfully promote the development of highly reliable domestic LLMs, adapted to Japan's legal systems and culture, thereby supporting the social implementation of administrative DX.
📋 Article Processing Timeline
- 📰 Published: March 31, 2026 at 22:00
- 🔍 Collected: April 1, 2026 at 13:39 (15h 39m after Published)
- 🤖 AI Analyzed: April 22, 2026 at 02:27 (492h 48m after Collected)
Visual Bank Inc. (Minato-ku, Tokyo; Representative Director and CEO: Masayuki Nagai) has announced that its AI training dataset solution, "Qlean Dataset," offered through its subsidiary Amanaimages Inc., will strengthen its data support system specifically tailored for administrative practices. This enhanced support targets companies selected for the Digital Agency's "Government AI (Gennai)" trial LLM program*.
This program is designed to powerfully support the adaptation of AI utilization in administrative practices to "factuality" and "Japan's unique context," which are fundamental to its success, through the provision of rights-cleared and high-purity datasets.
*Results of the public call for domestic Large Language Models (LLMs) to be trialed in Government AI: https://www.digital.go.jp/news/10d55c63-b3e1-42b9-9cc5-93a06943ae0e
## Qlean Dataset: The Foundation for Social Implementation of "Government AI"
In a society facing population decline, the utilization of AI is an urgent task for maintaining and strengthening public services. The "Government AI (Gennai)" project, led by the Digital Agency, is an extremely important initiative to ensure Japan's autonomy in AI and foster domestic LLMs optimized for administrative settings.
AI deployed in administrative practices requires "high reliability" that goes beyond general conversational abilities, demanding a deep understanding of Japan's legal systems, culture, customs, and the unique writing styles of public documents. Qlean Dataset, based on Japan's largest collection of rights-cleared data assets, addresses the government's stringent requirements from a data perspective. By providing solutions, it supports the process for each company's model to advance to the government's full-scale procurement phase. Beyond merely providing data, Qlean Dataset acts as a "strategic data partner" in the large-scale social implementation of administrative DX, offering powerful support for this initiative.
## Qlean Dataset's Data Quality and Value Meeting Government Procurement Standards
Qlean Dataset offers the following values to maximize the performance of models developed by the selected companies:
- **High Adaptability to Administrative Practices**
It supplies high-precision datasets for learning and evaluating public document-specific structures and highly specialized domains (e.g., legal, medical, social security), which cannot be achieved with publicly available internet data alone.
- **Objective Reliability and Factuality Verification**
To suppress hallucinations (plausible falsehoods generated by AI) and ensure factuality acceptable in administrative settings, it supports evidence-based model enhancement, such as building benchmarks using unlearned data.
- **Injection of Japan's Unique Values and Context**
By providing data that reflects the nuances of the Japanese language and Japan's cultural background and values, it accelerates the alignment towards "trusted AI" that resonates with users in Japan.
This program is designed to powerfully support the adaptation of AI utilization in administrative practices to "factuality" and "Japan's unique context," which are fundamental to its success, through the provision of rights-cleared and high-purity datasets.
*Results of the public call for domestic Large Language Models (LLMs) to be trialed in Government AI: https://www.digital.go.jp/news/10d55c63-b3e1-42b9-9cc5-93a06943ae0e
## Qlean Dataset: The Foundation for Social Implementation of "Government AI"
In a society facing population decline, the utilization of AI is an urgent task for maintaining and strengthening public services. The "Government AI (Gennai)" project, led by the Digital Agency, is an extremely important initiative to ensure Japan's autonomy in AI and foster domestic LLMs optimized for administrative settings.
AI deployed in administrative practices requires "high reliability" that goes beyond general conversational abilities, demanding a deep understanding of Japan's legal systems, culture, customs, and the unique writing styles of public documents. Qlean Dataset, based on Japan's largest collection of rights-cleared data assets, addresses the government's stringent requirements from a data perspective. By providing solutions, it supports the process for each company's model to advance to the government's full-scale procurement phase. Beyond merely providing data, Qlean Dataset acts as a "strategic data partner" in the large-scale social implementation of administrative DX, offering powerful support for this initiative.
## Qlean Dataset's Data Quality and Value Meeting Government Procurement Standards
Qlean Dataset offers the following values to maximize the performance of models developed by the selected companies:
- **High Adaptability to Administrative Practices**
It supplies high-precision datasets for learning and evaluating public document-specific structures and highly specialized domains (e.g., legal, medical, social security), which cannot be achieved with publicly available internet data alone.
- **Objective Reliability and Factuality Verification**
To suppress hallucinations (plausible falsehoods generated by AI) and ensure factuality acceptable in administrative settings, it supports evidence-based model enhancement, such as building benchmarks using unlearned data.
- **Injection of Japan's Unique Values and Context**
By providing data that reflects the nuances of the Japanese language and Japan's cultural background and values, it accelerates the alignment towards "trusted AI" that resonates with users in Japan.