Artificial Intelligence, Autonomy, and Human‐Machine Teams: Interdependence, Context, and Explainable AI: Editorial Introduction to the Special Articles on Context

Abstract


 
 
 Because in military situations, as well as for self-driving cars, information must be processed faster than humans can achieve, determination of context computationally, also known as situational assessment, is increasingly important. In this article, we introduce the topic of context, and we discuss what is known about the heretofore intractable research problem on the effects of interdependence, present in the best of human teams; we close by proposing that interdependence must be mastered mathematically to operate human-machine teams efficiently, to advance theory, and to make the machine actions directed by AI explainable to team members and society. The special topic articles in this issue and a subsequent issue of AI Magazine review ongoing mature research and operational programs that address context for human-machine teams.
 
 

The Interdependence of Autonomous Human-Machine Teams: The Entropy of Teams, But Not Individuals, Advances Science

“…The issue regarding limits of authority from checks and balances introduces discussion about the ethics of autonomous vehicles in use by the military in what are known as humans observing “on-the-loop” autonomous machines, but with humans not directing their decisions; these more fully autonomous operations carry significant risks. On the positive side, self-directed machines may save more lives since most accidents are caused by human error [ 1 ]. However, an editorial in the New York Times [ 117 ] articulated the concerns that AI systems can be hacked, suffer data breaches, and lose control to adversaries.…”

“…To set the stage, humans and machines are evolving into the age of autonomous human-machine teams (A-HMTs) at a time when the need exists to make ever-faster decisions than ordinary humans can process, which motivates the management of complex systems and A-HMTs with AI (e.g., from self-driving vehicles to reducing accidents to hypersonic missiles in combat; in [ 1 , 2 ]).…”

“…The human operator saw the pedestrian 1s early but hit the brakes 1s after impact. However, the Uber car did not update the context for its operator when it could have; its lack of interdependence with its partner made it a poor teammate [ 1 ].…”

“…In contrast, multitasking is the function of teams [ 8 , 44 ], making teamwork an emergent property that is unequal to the sum of a team’s individual contributors. Forming and operating a team requires multiple streams of communications (verbal, non-verbal) that include the constructive and destructive interference transmitted by interdependence; e.g., destructive interference among humans includes angry debate, even among the best teams [ 45 ]; constructive interference includes support for the behaviors that are chosen for a context, for which AI systems, especially machine learning, have been unable to manage so far [ 46 ], a problem that must be solved for effective and efficient human-machine teams (see our Special Issue in AI Magazine; [ 1 ]).…”

“…Smallman [ 57 ] proposed an adversarial approach for the US Navy as a way to reduce accidents at sea by subjecting decisions under uncertainty to a vote. Recently, the adversarial approach solved the Uber self-driving accident that killed a pedestrian in March 2018 [ 1 ]. At first, the Police Chief blamed the pedestrian, who was supposedly on drugs at the time.…”