Using AI-Supported Onboarding Systems in Video Games to Improve Player Experience

Abstract

Video games face the challenge of providing onboarding that motivates new players to engage with a game beyond their initial experience. Interactive media inherently influences players’ cognitive load during the learning process; video games must therefore determine a method of teaching new players game mechanics without exceeding their mental capacity for processing new information. Too much guidance can cause player frustration or boredom, while too little guidance can overwhelm. Instead of using restrictive onboarding methods, this thesis proposes that video games can use artificial intelligence systems that handle some in-game decisions to reduce new players’ cognitive load. To demonstrate this concept I designed and evaluated Joker, a turn-based strategy game with an AI-supported onboarding system that suggests an action on the player’s turn. I conducted a mixed-methods within-subjects study (n = 20) to examine the impact of AI-supported suggestions on new players’ cognitive load and to better understand the relationship between AI-supported onboarding systems and player experience. Results indicate that AI-supported suggestions successfully reduce players’ cognitive load, but that too low of a cognitive load negatively impacts players’ ability to learn from the AI-supported suggestions. Players primarily learn through lived game experience, and they strongly value interaction, agency, and personalization during the onboarding process. Future implementations of AI in onboarding should therefore ensure that AI-supported onboarding methods maintain a player’s ability to learn, and additionally use these dynamic systems to provide increased player control over the onboarding experience.