The term interpretability is primarily found in the fields of artificial intelligence, big data and smart data, and digital transformation. It describes the traceability and comprehensibility of complex digital systems and algorithms – in other words, how well we as humans can understand their decisions and functionalities.
In practice, interpretability means that we can understand why a piece of software or an AI has made a particular decision. For example, a bank uses artificial intelligence to make credit decisions. If this software denies a customer a loan, it is important for the bank, and also for the customer, to be able to understand how that outcome was reached. If the system is interpretable, it becomes transparent, for instance, that income or previous payment behaviour were the reasons.
Good interpretability helps to strengthen trust in digital systems, find errors and ensure fair decisions. Especially in sensitive business processes, understandable decisions play a central role so that we humans can truly trust the results of algorithms.
Deutsch 
English (UK)