Google Research

Are We Having Fun Yet? Designing for Fun in Artificial Intelligence That Is Multicultural and Multiplatform.

Human Agent Interaction Conference (HAI) (2019)

Abstract

Personality of Artificial Intelligence (A.I) is a vast and complex domain. Traits like trustworthiness and humanness have long been evaluated, but less so for traits like fun. In this study we evaluate fun in four conversational agents with a mix of menial tasks, games, and free play in the US and India on voice and text-based platforms. Our research explores user sentiment of fun in an A.I. personality, implications of fun personality on voice-based and text-based agents, and the perception of a fun A.I. personality in two culturally distinct markets. We used a mixed method approach to evaluate these objectives with 44 participants. Our results end with design principles for creating fun in artificial intelligence: 1) balance fun through conversational humor and goal-oriented actions, 2) create fun experiences through playful interactions, 3) convey fun through platform specific expressions like tone of voice or visual gestures and emojis, and 4) be inclusively fun by being culturally relevant so your conversational agent translates across borders.

Learn more about how we do research

We maintain a portfolio of research projects, providing individuals and teams the freedom to emphasize specific types of work