Comment by Stefan Buijsman

I propose in Sect. 5 that a lack of explainability may not give an immediate reason to reject the use of an AI system. It does, however, give us a strong reason to look for alternative setups that allow us to meet more of the procedural requirements ...

We should avoid applications of AI that largely replace decision-making (relegating humans to the position of checking outputs). We can, however, use AI in other roles even for high-risk decision making while conforming to all of the requirements set...