AAAI SSS-22 Symposium on Closing the Assessment Loop: Communicating Proficiency and Intent in Human-Robot Teaming
This work tackles the incremental problem of standardizing proficiency and intent communication in human-robot interaction, which is critical for improving team efficacy across AI, robotics, and HRI domains.
The symposium addresses the challenge of establishing standards for evaluating how robots communicate their proficiency and adapt to human intent in human-robot teams, highlighting the lack of agreed-upon methods despite prior evidence on its impact on human perception and decision-making.
The proposed symposium focuses understanding, modeling, and improving the efficacy of (a) communicating proficiency from human to robot and (b) communicating intent from a human to a robot. For example, how should a robot convey predicted ability on a new task? How should it report performance on a task that was just completed? How should a robot adapt its proficiency criteria based on human intentions and values? Communities in AI, robotics, HRI, and cognitive science have addressed related questions, but there are no agreed upon standards for evaluating proficiency and intent-based interactions. This is a pressing challenge for human-robot interaction for a variety of reasons. Prior work has shown that a robot that can assess its performance can alter human perception of the robot and decisions on control allocation. There is also significant evidence in robotics that accurately setting human expectations is critical, especially when proficiency is below human expectations. Moreover, proficiency assessment depends on context and intent, and a human teammate might increase or decrease performance standards, adapt tolerance for risk and uncertainty, demand predictive assessments that affect attention allocation, or otherwise reassess or adapt intent.