Designing for collaborative autonomy: updating user-centered design heuristics and evaluation methods

摘要

As we design automated and autonomous products that make increasingly sophisticatedrninferences and stronger interjections in a wider range of settings, it is increasingly critical tornconceptualize these products as cognitive agents, and not simply as passive tools. Our repertoirernof heuristics and techniques must expand to explicitly support not only a person’s ability to takernactions, but also to make sense of the world, determine the applicability of current and futurernplans, and select appropriate actions among many alternatives. These machine agents will also bernexpected to perform some or all of these functions themselves. Collectively, these attributes canrnbe thought of as facilitating collaborative autonomy, in which all agents in the system can expressrninitiative and cede authority based on their understanding of the world. However, product designrnis not the first discipline to face these problems or design these types of solutions. CognitivernSystems Engineering has been integrating and adding to the knowledge base in these areas forrnover 30 years. With some effort in translating their findings to our projects, we will be able tornaccelerate innovation and avoid the pitfalls and unintended consequences of previous attempts atrnincreasing inference and interjection.