Typhoon Bavi Preparedness: Taipei City to Launch AI Voice Care Service on July 8

(CNA, Taipei, July 7) With Typhoon Bavi potentially approaching Taiwan, the Taipei City Department of Social Welfare announced today that it will launch its "Disaster Prevention AI Voice Care Service" for the first time tomorrow. The service will reach 10,439 elderly individuals, is expected to save 10 minutes per case, and aims for a higher completion rate than the previous 50-60% achieved through manual calls.

The Taipei City Department of Social Welfare issued a press release explaining that the "Disaster Prevention AI Voice Care Service" project has undergone trial runs and system optimization since 2025. Following multiple scenario tests and feedback from elderly users, the voice recognition, conversational interaction, and needs reporting mechanisms have been refined and are ready for launch.

Zhao Jiahui, head of the Elderly Welfare Section at the Department of Social Welfare, told CNA that weather forecasts indicate the typhoon will have a significant impact on July 10th and 11th. Typically, calls to check on the elderly are made one to two days before natural disasters like typhoons, heatwaves, or cold snaps occur. Therefore, the service will be first implemented on July 8th and 9th, covering a total of 10,439 elderly individuals.

She mentioned that the current artificial intelligence (AI) voice outbound calling technology only supports Mandarin. After testing, the voice speed has been adjusted to make it easier for the elderly to listen. They will then respond by pressing telephone numbers based on the questions asked. Social workers have already informed the elderly in advance about the upcoming AI-powered care service to prevent them from mistaking it for a scam call and refusing to cooperate.

Zhao Jiahui analyzed the differences in using AI. In the past, if no one answered the phone, social workers had to redial after a period, consuming time and effort. Now, with AI handling these calls, it is estimated that 10 minutes of work time can be saved per case. Furthermore, the AI will compile the results, identifying individuals who are fine and those who require in-person visits, allowing social workers to follow up more efficiently and reducing their workload.

She anticipates that the completion rate will be higher than the 50-60% achieved with manual execution in the past. The execution results will be reviewed to continuously improve the system, making the care service more efficient and effective. (Edited by: Zhang Mingkun) 1150707

Stand with facts, your every donation is a force to protect press freedom

Download the CNA "One News" APP for real-time updates

Text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or used without authorization.