1. Challenges before implementation, and the reason for implementation

Spectee Pro notified us a serious disaster during the trial period

We started trialing Spectee Pro around the summer of 2018 and officially implemented it in April 2019. In terms of collecting information from citizens during disaster incidents, Toyota City primarily relies on phone calls. Our staff members also visit the disaster site and use video recordings taken with smartphones for situation assessment. However, the physical presence of staff at the disaster site can be burdensome. The impetus for our adoption of Spectee Pro came from a staff member who was temporarily assigned to the Cabinet Office and introduced the solution. At that time, we were intrigued by the AI feature that automatically extracts and notifies us of posts on social media related to disasters, which prompted us to introduce the system with a trial period. During the trial period, a significant landslide occurred in the mountainous areas of Toyota City, causing temporary road closures due to the extent of the damage. Spectee Pro was the first to notify us of this incident. In fact, Spectee Pro provided us with information about the situation before we received it from the road management authority. As mentioned earlier, our method of collecting information about the extent of damage was somewhat dependent on citizen phone calls and staff members visiting the site. This method sometimes fell short of capturing all the necessary information. During the trial period, we discovered that important information, including geolocation data through photos, was posted on social media. This realization led us to adopt Spectee Pro.

2. Utilization examples of Spectee Pro

Gathering the information of Aichi Prefecture which is provided through audio alerts, not only display

We utilize Spectee Pro to promptly grasp essential information such as disaster-related updates. During the initial stages of a disaster, information collection is often challenging. As a result, we focus on social media posts as one means of gathering information, especially during disasters like the potential Nankai Trough mega-earthquake. We also use it for gathering information during unforeseen events or accidents, as well as to understand the situation in other municipalities. Under normal circumstances, we don’t keep our eyes on the screen constantly. However, when there are incoming notifications, particularly when it involves audio alerts (by AI Announcer “Araki Yui”), we make it a point to go to our computer screens, checking the rainfall in different areas. The audio alerts are quite natural. During work hours, we can’t check our devices every time we receive a notification. Therefore, the feature of providing information through audio alerts, even when we’re not in front of our computers or not looking at our smartphones, is greatly appreciated. Our typical display is the homepage with the filter set to Aichi Prefecture. Even when there isn’t rainfall-related damage in Toyota City, we sometimes use Spectee Pro to observe the situation in other cities. We also use the search function to narrow down results based on date and location, and occasionally review past disaster-related incidents.

3. Effects after installing Spectee Pro

Concerns and anxieties about using SNS significantly reduced Since introducing Spectee Pro, our awareness of utilizing SNS posts has evolved. We served as the Chairman of the Core City Mayors’ Association, and during that time, when there was a significant disaster in Nagano Prefecture, we were involved in dispatching personnel and supplies, as well as coordinating with other core cities. When we were about to head to Nagano Prefecture, using Spectee Pro to understand the local situation was incredibly helpful. The ability to monitor the situation in other municipalities beyond Toyota City and Aichi Prefecture is of significant importance. Additionally, even before the Chikuma River breached, there were numerous posts related to heavy rain and rising river levels, which prompted us to consider the utilization of information even before a disaster occurs. After the introduction of Spectee Pro, the added value of AI-driven information extraction became apparent. This led to increased attention within the Disaster Response Headquarters regarding the use of SNS. Previously, there were concerns and anxieties about using SNS, but they were significantly reduced, and we became more proactive in utilizing it. Comparing the situation to the past when we handled disaster-related posts from SNS manually and had to assess their credibility, it is rather unsettling. Planning to establish a system for collecting information from SNS However, there is still a dilemma when it comes to handling SNS posts effectively and proactively. For example, there is a desire to use Spectee Pro in the fire department, but when it concerns human lives, deploying based on SNS posts is challenging. There is a concern that by encouraging people to use SNS as an alternative to calling 119 (emergency number), it could create misconceptions. In Toyota City, the challenge lies in establishing a structure to connect SNS information to disaster response. Starting from April next year, we plan to establish a system for collecting information from SNS proactively, rather than passively waiting for phone calls.

4. Future outlook

In terms of the Disaster Preparedness Division, we have diversified our means of information dissemination, such as introducing disaster radios and a registered email system, moving away from solely relying on disaster administrative radios. Looking ahead, we are exploring the possibility of automating the information intake process. Additionally, we are considering more ambitious ideas, such as using drones for delivering supplies. Toyota City has a significant foreign population. Disseminating disaster information to foreign residents is challenging, and, conversely, understanding foreign language posts on SNS is also a current challenge. We believe that organizing SNS posts by foreign residents or delivering disaster information via automated translation for foreign language speakers would allow us to provide more detailed support. In the future, we envision using AI and satellites to analyze satellite images for detecting discrepancies between pre-disaster and post-disaster conditions. This would enable us to swiftly assess the extent of damage in the event of disasters like landslides and flooding. Such information would be crucial for concentrating personnel allocation in administrative tasks, such as issuing disaster victim certificates. Given the opportunity, we would like to pursue these initiatives.