Researchers Test How to Make AI Robots Explain Their Decisions in Real Time
Scientists are experimenting with ways to make social robots transparent about how they make decisions during conversations, a critical step before deploying them in schools and hospitals. The work addresses a growing business and liability concern: as AI systems make decisions affecting students and patients, organizations need ways to prove—and explain—what the robot actually did.
Originaltitel: Speaking Transparently: Social Robots in Educational Settings
<p>The recent surge in popularity of Large Language Models, known for their inherent opacity, has increased the interest in fostering transparency in technology designed for human interaction. This concern is equally prevalent in the development of Social Robots, particularly when these are designed to engage in critical areas of our society, such as education or healthcare. In this paper we propose an experiment to investigate how users can be made aware of the automated decision processes when interacting in a discussion with a social robot. Our main objective is to assess the effectiveness of verbal expressions in fostering transparency within groups of individuals as they engage with a robot. We describe the proposed interactive settings, system design, and our approach to enhance the transparency in a robot's decision-making process for multi-party interactions.</p>