Presentation

· Contributors · Organizations · Search Program · Flagged · Happening Now

Function Allocation for Responsible Artificial Intelligence (AI): How do We Allocate Responsibility?

DescriptionTraditionally, Function Allocation has concentrated on comparisons between humans and machines in terms of their strengths and weaknesses, the most well-known of these being the Fitts MABA-MABA list (Fitts et al., 1951). With the advent of more advanced technologies such as artificial intelligence (AI) and robotics there is a need to move beyond conceptions of FA as discrete allocations (e.g., tasks, functions, goals, and outcomes) and the need to move towards the view that new forms of dynamic interdependencies will create shared, collective entities involving humans and machines. Interdependencies play a large part in recent approaches to FA. For example, Coactive Design defines the increasingly sophisticated roles that people and robots play as interdependent teammates engaged in joint activities. At the moment we have no established guidance in the form of HFE-oriented criteria or structured questions regarding how we might allocate ‘responsibility’ between humans, machines, and numerous forms of new technology (e.g., robots, drones). Likewise, the interdependencies which might exist in human-machine interaction are only starting to emerge (e.g., robotic systems which have the ability to ‘trust’ their human operators). The panel will aim to discuss these issues amongst others that arise as a result od developments in new technology.

Moderator