Assessing the artificially intelligent workplace: an ethical framework for evaluating experimental technologies in workplace settings.

Experimental technologies, including AI and robots, are revolutionising many types of work. For example, the logistics warehouse sector is witnessing a wave of new technologies, such as automated picking tools, collaborative robots and exoskeletons, affecting jobs and employees. Notably, it is not always possible to predict the effects of such new technologies, since they have inherent uncertainties and unintended consequences. Hence, their introduction into workplaces can be conceived as a social experiment. This paper aims to sketch a set of ethical guidelines for introducing experimental technologies into workplaces. It builds on Van de Poel's general framework for assessing new experimental technologies and translates that framework into a more specific context of work. We discuss its five principles: non-maleficence, beneficence, responsibility, autonomy, and justice. Each of these principles is applied to workplaces in general, and specifically to the logistics warehouse setting as a case study. A particular focus in our discussion is put on the distinctive potential harms and goods of work.