Safety Issues in Human-Machine Collaboration and Possible Countermeasures

Abstract

Autonomous machines are more and more capable of executing complex tasks with the support of intelligent algorithms, and they are deploying rapidly at an unprecedented pace. In the meanwhile human-machine teaming is promising to accomplish more and more challenging tasks by integrating strengths and avoiding weaknesses from both sides. However, due to imperfections from both human and machine sides and their interactions, potential safety issues should be considered in advance so that researchers and engineers could prevent or tackle those issues with preparation and make the human-machine system safer and more successful. In this paper, we proposed a framework under the context of human-machine (algorithm) collaboration, and we addressed possible safety issues within and out of the human-machine system. We classified those safety issues into internal safety issues representing the safety issues within the human-machine system and external safety issues representing safety issues out of the human-machine system to organizational and societal levels. To tackle those safety issues, under this proposed framework, we listed possible countermeasures according to the literature so that we could provide pedals to control the autonomous agents and human-machine teaming and enable safer human- machine collaboration in the future.