Public Acceptance of AI Agent with Accountability

Abstract

We propose to develop an AI agent with accountability as a way to spread AI technology into fields related to human life, body, and property safety, such as medical care, autonomous driving, and finance. When discussing who has accountability, AI developers, AI sellers, AI owners, etc. are often considered. In contrast, this paper discusses whether AI agents themselves can take accountability and whether such agents can be accepted by society. We prototyped a video recommendation agent that is accountable for the task and made a preliminary experiment to evaluate the acceptability of the AI agent in everyday life. Based on the experimental results, future research issues and future prospects are discussed.