²⁸For example, the General Data Protection Regulation (GDPR) of the European Union imposes a general “right for explanation” for almost any decision made by an algorithm on an individual (Goodman and Flaxman, 2017). One reason for such a requirement is to make sure that the AI did not discriminate applicants based on gender, race, or similar characteristics—an objective called “fair AI”. Another reason is that the AI might not make the final decision, but could be used as a support for a human decision-maker, such as a medical doctor: The human decision-maker would greatly benefit from understanding why the AI came to its conclusion. One more reason for making AI easy to interpret is that understanding how an AI works makes it easier to evaluate its potential safety hazards, and develop AI that is safe.