Rebalancing Worker Initiative and AI Initiative in Future Work: Four Task Dimensions

Abstract

Organizations have recently begun to deploy conversational task assistants that collaborate with knowledge workers to partially automate their work tasks. These assistants evolved out of business robotic process automation (RPA) tools and are becoming more intelligent: users can initiate task sequences through natural language, and the system can orchestrate those tasks if they have not previously been defined. As these tools become more automated, system designers tend to optimize overall process efficiency, but at the cost of shifting agency away from workers. Particularly in high stakes work environments, this shift raises questions of how to re-delegate agency such that workers feel sufficiently in control of automated tasks. We explored this through two studies comprised of interviews and co-design activities with knowledge workers and identified four task dimensions along which their automation and interaction preferences vary: process consequence, social consequence, task familiarity, and task complexity. These dimensions are useful for understanding when, why, and how to delegate agency between workers and conversational task assistants.