April 2019 to March 2020

Izumi Harada (Executive Fellow, Institute for International Socio-Economic Studies)
Yusuke Koizumi (Chief Fellow, Institute for International Socio-Economic Studies)

(1) Domestic and international issues in AI ethics and privacy
Various issues with Big Data analysis using AI have been identified in areas of privacy and other human rights. These include issues such as inference of special care-required personal information (or sensitive personal information) by profiling, furthering social discrimination due to biases, and transparency of algorithms. Government agencies in Japan and other countries have published guidelines to deal with such issues and indicating general principles such as fairness, accountability and transparency, but not concrete measures that a company must put into practice, so further study and discussion are needed. 

 (2) Services using facial recognition technology and related regulatory trends
As a way to confront GAFA and other huge platform enterprises, initiatives have begun in Japan for an integrated service among multiple enterprises that will increase portability of personal data, such as Personal Data Services/Stores (PDS), and Personal Data Trust Banks. Three common challenges with such services are monetization for the operator, incentives for users, and the ID used to identify users. Services on the Internet can use a textual ID or email address, but services in the real world have required users to present a card or their smartphone. Replacing such real world authentication of individuals with facial recognition or other biometric method has begun at airports and some stores, but there is not yet sufficient consensus in society, regarding issues such as privacy, to expand its use more broadly. Further study of domestic and international trends regarding services using facial recognition and of policies needed to gain consensus in society are needed.

In this research, we conducted a literature survey on these two themes, conducted field studies (in the UK and France), made recommendations, and published the results inside and outside the company.