“traditional” ML techniques vs AI for explainability and reproducibility:

[StatsRhian]

a session about the tradeoffs: learn hard thing, gain skills and understand what you're doing/what the outputs mean, versus much greater user simplicity but potentially much lower certainty in results (without necessarily being able to even quantify the error)