Distrust for Human Judgment: Aspiration for Automated Decision in Japanese Nuclear Risk Governance

• In Event: Automation, Skill and Identity in an Age of AI - I

Abstract

The author co-authored the analysis of sociotechnical interactions on a particular belief in the context of nuclear risk governance in Japan, which is shared not only among the policy-makers and technical experts, but also by citizen stakeholders: “calculation should automate decision and prevent regrettable consequences.” It was case studies of the probabilistic risk assessment (PRA) and the System for Prediction of Environmental Emergency Dose Information (SPEEDI) and presented at the last 4S meeting. He this time expands the analysis by comparing those cases with the other case of nuclear risk management: high-level radioactive waste management.

The key concepts of those sub-field in nuclear risk management have foreign origin, have developed and/or elaborated internationally, and finally accepted in Japanese context. In the historical process of the policy discussions, however, the use of quantitative risk calculation technic has been interpreted, transformed and implemented in particular manner: too strong dependence on and too high expectation to the reliability of it. It has been endorsed by the belief mentioned above. Almost all stakeholders have been skeptical about the feasibility and trustworthiness of human judgment and aspired to have automated decision directly supported by advanced computer-based calculation of the risks.

This paper critically examines it by connecting the sociological concept “structural disaster” (Matsumoto 2013) and another STS interpretive framework “sociotechnical imaginary” (Jasanoff and Kim 2009, 2015). By taking the advantage of this theoretical connection, it aims to expand our insights on the repeating and ubiquitous occurrence of similar failure trajectories of nuclear risk management policy.

Author

• Kohta Juraku, Tokyo Denki University