Human Agent Collaboration

Human-agent collaboration (HAC) research focuses on designing AI systems that effectively partner with humans to achieve shared goals, emphasizing seamless interaction and mutual understanding. Current efforts concentrate on improving explainability of AI decision-making, developing efficient online adaptation methods (like bootstrapped logistic regression) to align AI behavior with human preferences, and designing communication protocols to facilitate effective teamwork, particularly in complex scenarios such as online gaming and customer service. These advancements are crucial for building trustworthy and beneficial AI systems, impacting fields ranging from customer support to complex task solving where human expertise and AI capabilities can synergistically enhance performance.