SatisfAI: A Serious Tabletop Game to Reveal Human-AI Interaction Dynamics

Interactions between humans and AI-enabled systems are occurring more frequently, often at times unbeknownst to the human end users. As AI-enabled systems become more pervasive and advanced in helping humans make decisions, there is a need to ensure the systems are designed around human needs and for the end users to have a general understanding of the benefits and limitations of AI. Humans who understand both the benefits and limitations of AI systems may then make more informed decisions that maximize the satisfaction of their own needs. Toward this end, we have developed a game based on Max-Neef's Fundamental Human Needs Scale to inform human-AI interaction design. We have designed and prototyped the game so that it may serve as a data collection and user research tool. The tool presents various human-AI interaction scenarios to players in an effort to reveal the level of agency humans are comfortable ceding to an AI-enabled system, based on their personal needs that are satisfied by that system. Through the design mechanics and gameplay that result from real people playing the game, insights may be gathered regarding human-AI interaction dynamics, which may then be used to inform the design of future AI systems.