In chapter 6, we discuss the second important aspect of implementing an AI system: testing and validating AI. When an AI system is good enough to proceed to implementation is not only a technical issue, but also a management issue. Based on the cases, we discuss that an AI system must function properly technically, must comply with relevant laws and regulations, and should be in line with ethical guidelines. However, we show that in order to achieve a properly functioning AI system, it is just as important to test and validate the system, taking into account the work processes and the 'social' requirements of the user and the organization. We discuss that testing and validating requires extra attention, because how AI systems come to conclusions is black-boxed and therefore difficult to explain. This also complicates the question of who is responsible regarding final decisions.
You are not authenticated to view the full text of this chapter or article.
Get access to the full article by using one of the access options below.