Context-aware Agents based on Psychological Archetypes for Teamwork

As artificial agents become ubiquitous with the advancements of Artificial Intelligence (AI), creating effective, viable artificial teammates has become increasingly important. Multiple research studies have attempted to personify agents endowing them with different archetypes observed in human psychology theories with the aim of creating realistic, predictable and believable agents. However, when these agents are exposed to other agents (both artificial and human), the archetypal qualities should be amenable to create socially believable and socially intelligent agents. This paper presents a generic framework to model personified archetypes of agents. The framework provides a flexible platform that can accommodate the behavioral changes of an agent influenced by many contextual factors. The proposed framework will drive better designing of effective believable synthetic agents/characters and more user-friendly virtual assistants customized to a human's personality.