Evaluation of Performance

Abstract

Agents evaluate their performances to assess progress, learn, and improve. In doing so, they refer to criteria of various kinds. Some criteria are deeply encoded in mental models, organizational procedures, or cultural norms and logics, while other evaluative criteria are adaptive and may upregulate or downregulate, depending on the agent’s goals, expectations, and context. Here, too, digitalization is transformative. Artificial agents bring unprecedented power to the evaluation of performance, including the rapid intra-cyclical evaluation of ongoing processes. These mechanisms support feedforward guidance in real time. Therefore, when human and artificial agents combine in the evaluation of augmented performance, they face additional risks. Artificial evaluative processing could be fast and precise, while at the same time, human evaluation may be relatively sluggish and imprecise. Overall evaluations of performance could be distorted and dysfunctional.