Yamaguchi City and Core Corp. Conduct Pilot Project on Generative AI for Disaster Call Response

## Yamaguchi City and Core Corp. Conduct Pilot Project on Generative AI for Disaster Call Response

Core Corporation (HQ: Setagaya-ku, Tokyo; CEO: Koji Yokoyama; hereinafter 'Core') has partnered with the Disaster Management Division of Yamaguchi City, Yamaguchi Prefecture, to conduct a pilot project on a phone reception support system utilizing generative AI, aimed at improving efficiency and speeding up initial response during disasters.

### 1. Background and Challenges
During disasters, municipal emergency headquarters face an overwhelming volume of calls, requiring significant resources. In municipalities with high staff rotation, new or support personnel often handle initial response, leading to inconsistency in quality. The field urgently requires solutions for preventing information omissions, ensuring data accuracy, facilitating rapid information sharing, and reducing staff stress.

### 2. Pilot Project (PoC) Overview
The PoC was conducted to evaluate the system for future full-scale implementation:
- **Duration**: 3 days (February 3–4 and February 12, 2026)
- **Location**: Disaster Management Division, Yamaguchi City, Yamaguchi Prefecture
- **Verification Points**:
1. Accuracy of voice recognition and transcription
2. Automated generation of disaster reports
3. Advice and reminders for input omissions
4. Multi-lingual transcription and automated report generation

### 3. Key Results
The following results were confirmed:
- **Transcription Accuracy**: Verified that it functions well even with dialects and onomatopoeia (average score: 3.9/5).
- **Automated Report Generation**: Instant extraction of names, contact info, and addresses from call contents, with a potential 70% reduction in report creation time.
- **Support Functionality**: Confirmed effectiveness of input omission reminders, with actionable UI improvements identified (average score: 3.8/5).
- **Multi-lingual Support**: Successfully tested Chinese, French, Korean, and Spanish transcription and report generation.

### 4. Overall Evaluation
85% of participants gave the highest possible rating, with a majority favoring practical implementation. Mr. Ota from the Yamaguchi City Disaster Management Division noted that the system compensates for variations in staff experience and significantly contributes to the quality and speed of initial response.

### 5. Future Outlook
Moving forward, Core aims to evolve the system beyond a transcription tool into a robust 'decision-support tool' by enhancing features such as map integration and recognition of localized terms, strengthening situational awareness and information sharing during disasters.