Last September (2020), Taro Kono, Minister of State for Administrative Reform, set up an “Administrative Reform Reference Box (Stove-piped 110)” on his website to solicit opinions, suggestions, and complaints from the public about the adverse effects of stove-piped administration. Many people may remember that over 4,000 opinions flooded in on the first day of its opening and the following day, and that it made the news when it stopped accepting them the next day. YOSHINA analyzed over 4,000 opinions at this time. The policy was for Minister Kono to read through all of them, but he was having difficulty with the sheer number. Initially, he used his breaks and other time to conduct a thorough examination of the opinions in an analog manner, but once he started analyzing YOSHINA, he was able to complete the process in just five minutes. The opinions received were sorted into a total of 30 categories (25 individual policies + 5 general comments), and Kawahara reported back to the minister. Of these, the top 15 individual policies that received the most opinions were presented in detail and received lively questions from the minister about the results of the analysis. “The essence buried in over 4,000 opinions. The true meaning hidden in the vast forest of words. It would be an impossible task, both in terms of time and quality control, to search these out through human work. With “YOSHINA,” all opinions are categorized by topic and arranged in order of frequency of occurrence, and from each category, “common word groups” and “noteworthy word groups that provide hints for awareness” are identified, and deep investigation is automatic. We presented our analysis based on our policy of “Separating is understanding,” and the minister seemed convinced by our analysis. If this “to divide is to understand” is the fundamental vision, “YOSHINA” sees “democratization of AI” as an important mission. In the presentation that wowed Minister Kono, there are no special tasks that only a group of AI experts can perform in order to use the platform. The process is quite simple. (1) Drag and drop the text data you want to analyze onto the screen. (2) Follow the navigation and check the configuration items. Just by following these two steps, the analysis results are displayed on an easy-to-understand screen with graphs. Even for companies without knowledge or resources related to AI, there are no barriers to introducing the system.
- I think the essence of broad listening is “technical assistance to ensure that the opinions of a large number of people are taken into account as much as possible” emphasis added. From that perspective, I think this case (even though it wasn’t called that at the time) can be called an example of the political use of broad listening.
This page is auto-translated from /nishio/YOSHINA using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.