In the past few months, a new debate started about the so-called explainable AI, an AI which isn't a sort of black box (where we understand only the underlying mathematical principles) and whose actions or decisions can be easily understood by humans. Criticism has been also been made (in general for AI, but in particular DL) about the generated results from models not being compliant with GDPR (short for General Data Protection Regulation): https://ec.europa.eu/commission/priorities/justice-and-fundamental-rights/data-protection/2018-reform-eu-data-protection-rules_en for data related to EU citizens, or other data regulations that will probably be defined next in other parts of the world, which require the right to an explanation to prevent discriminatory effects based on different factors.
While this is a real hot and not negligible topic, and since...
