Social Interaction with Autonomous Systems

Autonomous systems can adapt to change and spontaneously develop new behaviour by learning from experience. This makes them crucially different from earlier generations of artificial systems the behavior of which is programmed and predictable. Social autonomous systems display yet another crucial difference, designed to work on similar tasks as humans together with humans, as partners. Previously, autonomous systems have at most worked close to or side-by-side with humans.

Social autonomous systems, SAS, very soon will be a common yet disruptive element in daily life, at work, at home, in school, and so on. For this to function adequately and be unproblematic requires the development of novel models of human-machine interaction, and designing new ways of interacting that respect privacy and safety, feel comfortable for human users, are efficient and effective, and much more. So far nobody knows what other values may be significant, or how such interaction on a daily basis will affect human comportment and productivity, mental health, the experience of meaningfulness, social trust, norms and habits, nor how the interaction could be designed to maximize human potentials and minimize harm both on the societal and personal levels.