Joint activity, as we define it, is a mutually interdependent social endeavor that requires sufficient predictability among participating parties to enable coordination. Coordination, in turn, sometimes requires the parties to appraise the state of progress of their activities so that, if necessary, they can adjust their actions to meet coordination needs and communicate their status to others as appropriate. A significant impediment as yet precluding the full participation of automation in joint activity with people is its inability to sense and communicate aspects of its state that would allow other participants to meaningfully assess progress toward (or anticipate failure with respect to) mutual objectives. In the current article, we address various issues associated with “progress appraisal” and the challenges it poses for human-machine systems. We point to promising directions for future work.
Paul J. Feltovich, Jeffrey M. Bradshaw, William J.