The Women in AI speaker series, a collaboration between the Schwartz Reisman Institute for Technology and Society and Deloitte, welcomes Arisa Ema, an associate professor at the University of Tokyo whose research investigates the benefits and risks of artificial intelligence (AI) systems through interdisciplinary approaches. Ema is a co-founder of the Acceptable Intelligence with Responsibility study group, a board member of the Japan Deep Learning Association, and was a member of the Council for Social Principles of Human-Centric AI, which released “Social Principles of Human-Centric AI” in 2019.
In this talk, Ema will survey how to translate values and principles established for ethical AI development into practice in the context of industry applications, foregrounding the importance of robust governance ecosystems that ensure accountability and risk management for deployed AI systems.
Talk title:
“From principles to practices: Building an AI governance ecosystem”
Abstract:
Since 2016, an increasingly robust discussion around what principles ought to guide the development of AI systems has produced a general consensus on what values are important. In this talk, I will consider the relationship between the principles required for AI development and their practical implementation. While the title of this talk references “principles to practices,” in reality the arrow of development between these areas is mutually inclusive and beneficial.
While there are many ways to put AI principles into practice, I will focus on best practices in the context of AI governance ecosystems. Within this context, it is important to consider the industry structure that surrounds the development of AI systems. Compared with platform companies, Japan has many business-to-business (B2B) companies; in the development of AI services and systems, data acquisition, AI model development, and service provision tend to be done by different actors. This makes the issue of who is responsible when an accident occurs in the course of providing services more challenging. Consequently, it is important to establish a clear governance ecosystem that includes such features as auditing, insurance, whistle-blowing, and accident investigation, among other measures.
Consideration among various stakeholders of the risk factors posed by AI systems requires an additional governance framework. In the latter part of my presentation, I introduce a risk chain model (RCModel) created by the University of Tokyo, and discuss how to mitigate and control risks.
Recommended readings:
Study Group on AI Governance and its Evaluation, “AI Governance Ecosystem: Who Manages and Evaluates AI?” and “AI Governance Ecosystem: Trusted AI with Industrial Structure,” Japan Deep Learning Association, 2022.
AI Service and Risk Coordination Study Group, “How to use Risk Chain Model,” Institute for Future Initiatives, 2022.
Arisa Ema, “Visions, Issues, and Measures to Discuss AI and Society,” 2021.
Arisa Ema, “Oxford Conference Challenge: Realizing Diversity, Inclusion, and Sustainability,” Nikkei XTECH, 2019.
Arisa Ema, “Task Dividing Workshop: Co-create Work and Life with Artificial Intelligence,” ER No. 4, 2017.
About Arisa Ema
Arisa Ema is an associate professor at the University of Tokyo, a visiting researcher at RIKEN Center for Advanced Intelligence Project in Japan, and a research support advisor at AIST Department of Information Technology and Human Factors. Ema is a researcher in science and technology studies, whose primary interest is to investigate the benefits and risks of artificial intelligence through interdisciplinary approaches. She is a co-founder of the Acceptable Intelligence with Responsibility study group established in 2014, which seeks to address emerging issues and relationships between AI and society. Ema is a board member of the Japan Deep Learning Association, and was also a member of the Council for Social Principles of Human-Centric AI, which released “Social Principles of Human-Centric AI” in 2019. Ema obtained her PhD from the University of Tokyo and previously held a position as assistant professor at the Hakubi Center for Advanced Research, Kyoto University. In 2021, she received the University of Tokyo’s Excellent Young Researcher Award.
About Women in AI
Women in AI is a six-part virtual speaker and mentorship series developed by the Schwartz Reisman Institute for Technology and Society in collaboration with Deloitte that connects a global audience with a diverse group of female thought leaders in the field of AI research. The Women in AI series convenes leading female AI researchers to share knowledge and mentorship opportunities through seminar events that promote opportunities for women across the technology sector. Participants will explore how women are leading the development of new technologies and approaches, and investigating emerging trends across the sector. The series provides insights into cutting-edge research, bold points of views, and help business and community leaders elevate diverse voices while promoting opportunities for women to share their perspectives.
Arisa Ema