How does the boundary change between being a user of AI tools and needing to be a subject matter expert - under the hood? Where does one need transparency (auditability) under the AI algorithm hood when they are accountable for the result of using the tool? Thank you,