Mercari, Inc. (“Mercari”) research and development organization Mercari R4D (“R4D”) and Osaka University Research Center on Ethical, Legal and Social Issues (“Osaka University ELSI Center”) have unveiled a paper addressing their joint research into ethical, legal and social issues (“ELSI”) around AI. The paper, titled “Addressing trade-offs in co-designing principles for ethical AI: perspectives from an industry-academia collaboration,” is available for download.
Click to download the full text of the paper: https://doi.org/10.1007/s43681-024-00477-8
Abstract
The paper examines the analysis of minutes taken at approximately 20 meetings that took place between Osaka University ELSI Center researchers and Mercari Group members in 2021 and 2022. The aim of the meetings was to formulate unique organizational AI ethics principles for Mercari.
ELSI Center researchers first analyzed the content of the meeting minutes and sorted through the opinions and issues related to discussions on leveraging AI. This allowed them to isolate four trade-offs that the meeting participants faced.
The four trade-offs
- We should leverage the benefits of AI systems to the maximum extent possible and utilize data-driven algorithms, but we must not allow this to lead to the reproduction of discrimination.
- We will advance automation and introduce AI-driven decision-making; however, we must also maintain human-focused systems.
- We must protect the company’s proprietary and personal information, but we must also ensure there is transparency and accountability in the development and usage of AI systems.
- We use AI to provide recommendations and promote meaningful transactions while ensuring that we maintain an open market based on free participation.
The paper, which is based on the analyses and discussions of these trade-offs, is a collaboration between academia and industry that examines these complex issues and presents useful approaches to dealing with them.
Furthermore, based on the analysis results, the approach of our organization’s leadership toward AI ethics has highlighted two topics for future discussion: the need for further research to understand the long-term effects of AI ethics principles on our organization and society, and the need for more reports on the exact processes private organizations use to formulate these AI ethics principles.
The ELSI Center has also publicized the release of the paper on their website.
https://elsi.osaka-u.ac.jp/research/2855
Comments from the authors
- Amelia KATIRAI (Specially-Appointed Assistant Professor, Research Center on Ethical, Legal, and Social Issues, Osaka University)
- Many companies wonder how they should approach AI usage, and the development of ethics principles that are unique to each company has been raised as an answer. However, one issue with this is that a lot of the time, the details of these processes are not published, and there are still very few cases where the details of these processes are analyzed academically. Thanks to Mercari’s open stance on this issue, we were able to analyze and publish a process used for formulating ethics principles, and we have also been able to report on Mercari’s unique way of approaching the many trade-offs associated with the usage of ethical AI. I would like to express my gratitude to all the members who contributed to this research, starting with the Mercari AI team.
- Yusuke NAGATO (Specially-Appointed Assistant Professor, Research Center on Ethical, Legal, and Social Issues, Osaka University)
- I attended about 20 meetings, which collectively took a considerable amount of time and focused on formulating AI ethics principles. I had the chance to walk apace with Mercari members as they set off on the ambitious trek of formulating ethics principles almost entirely from scratch. This alone taught me a lot. Going forward, I would like to continue to advance research into AI ethics while following the activities involving Mercari Group’s AI principles.
Comment from Mari Inoue, Mercari R4D representative in charge
The results of our two-year-long joint research conducted with the Osaka University ELSI Center regarding AI ethics have now been selected and published as an academic paper.
The paper analyzes how Mercari approached the ethical issues pertaining to AI usage, the concrete processes involved, and the trade-offs that we faced. The fact that the paper affords Mercari’s initiatives for AI ethics the opportunity to be evaluated in both academic and international arenas is deeply meaningful.
Going forward, we will continue to proactively promote partnerships between industry and academia in the humanities, sciences, and social sciences, targeting Mercari’s diverse research areas and the company’s activities.
For inquiries regarding this research, please use the contact form on the R4D website.