Responsible AI Research Group

Welcome to the Responsible Artificial Intelligence (RAI) Research Group at the Kelley School of Business. Our mission is to advance research and education in responsible AI, focusing on the ethical design, deployment, and use of human-centric AI conversational agents. For over seven years, our faculty has explored how these technologies can create value for both customers and employees—highlighted by our recent feature in the Harvard Business Review. Our initiative aims to further expand research on RAI topics and provide educational programs for undergraduates, graduate students, and business leaders.

The Responsible AI Research Group is committed to exploring and addressing the ethical, social, and technical challenges of AI technologies. Our team of experts works on diverse projects, ranging from bias mitigation to enhancing AI transparency. Together, we aim to build AI systems that are safe, fair, and aligned with societal values.

Research Framework for Human-Conversational AI Interaction

As companies increasingly adopt conversational AI to stay at the forefront of innovation, they often face challenges in determining the appropriate use and user perception of these technologies. Our research framework offers a straightforward approach, focusing on three essential aspects of conversational AI:

• It is human: Conversational AI can engage in natural dialogue, collaborate with users, and provide a sense of assistance rather than acting as a passive tool.

• It is not human: While it eliminates challenges inherent to human interaction, AI lacks genuine human understanding. Users may react more strongly to AI errors, such as prediction mistakes, which can erode trust.

• It is superhuman: AI operates without fatigue, offers 24/7 service, scales quickly, and can access vast data instantly, ensuring comprehensive support across domains.

This framework helps researchers and practitioners ask key questions:

• What context is the AI used in, and what outcomes are important?
• What benchmarks and capabilities should we prioritize, such as interactivity, humanness, or scalability?

With over a dozen research projects and nearly two dozen studies completed, we are leading the development of practical, theoretical insights that benefit firms deploying Responsible AI.