The University of Tokyo, Dentsu Public Relations Inc., and Hotto Link reveal conclusions
of legislator survey on AI-identified social issues

Tokyo, Japan (July 27, 2018) – The results of a joint AI research project and legislator survey by the University of Tokyo, the Corporate Communication Strategic Studies Institute of Dentsu Public Relations Inc., and Hotto Link Inc. (TYO: 3680) have been announced, with implications for the shaping of public consensus and public policy.

The researchers used artificial intelligence (AI) to analyze a big data sample of Twitter posts from July 2017 and identify issues relating to social innovation, defined as “issues which, if solved, would bring major changes and improvements to society, the economy, and people’s lifestyles.” The program found a total of eighteen such issues, and these formed the basis of a survey conducted among legislators belonging to the National Diet of Japan. Respondents answered questions about their awareness of the issues identified and their perceived need for the relevant information required to address these challenges, as well as what kind of information is necessary when considering policies on innovation and what sources of information they currently use.

One notable finding was that unlike topics which are prominently discussed in particular communities, issues that are commented on less frequently and are more broadly scattered across social media are not well recognized by legislators, who tend not to notice these voices. Additionally, legislators indicated that they felt little need for information of the kind that would be critical to addressing these types of latent, broadly dispersed issues.

When asked about their approach to innovation-focused public policy, while approximately 80 percent of legislator respondents expressed a desire for fact-based evidence, a slightly greater proportion felt that the necessary information was lacking at the present time.

Overall, the results suggest that in order to create the right conditions for addressing issues relating to social innovation, it will be crucial to pick up on the public’s concerns through the analysis of big data and other sources, and further to present these issues to the government together with persuasive supporting evidence.

■ Research Overview
  • Research conducted by:

      – Sakata & Mori Laboratory, Department of Technology Management for Innovation, School of Engineering, the University of Tokyo
      – Corporate Communication Strategic Studies Institute (Dentsu Public Relations Inc.)
      – Hotto Link Inc.

1) AI for identifying issues relating to social innovation/big data analysis

  • Research overview: Two types of AI were developed on the basis of Twitter data and the government document Investments for the Future Strategy 2017. One program identified communities on social media, while the other identified issues relating to social innovation. The AI analysis revealed eighteen issues, which were then categorized into four groups according to 1) the number of comments and 2) topic entropy, or the concentration of comments within a particular community on Twitter.
  • Data analyzed: Random sample of several tens of millions of Twitter posts in July 2017

2) National Diet member survey

  • Survey method: The survey was sent to all National Diet members (both upper and lower house) with the help of an existing parliamentary group formed to promote a digital society. Responses were obtained via paper survey forms or online.
  • Survey period: Dec. 9, 2017-Apr. 8, 2018
  • Responses collected: 64 (43 upper house and 21 lower house members)