I would say the debate currently going on is less about "running code on your own machine" and more about "making sure the thing your are replacing at least a portion of your labor force with is at least somewhat dependable and those who benefit from the replacement are still responsible".
I think management is putting too much hope into this, any negative outcome from replacing a human with AI might result in liabilities surpassing the savings. Air Canada's chatbot was decided just a year ago and I'm sure the hallucinating AI chatbot, from development to legal fees, cost the airline more money than they saved in their call-center.
