> Deleting the system doesn’t undo damage from erroneous decision-making

How is this different from a human CEO? Removing them from the position doesn't erase the impact of the decisions that led to them being removed.

What happens if the board wants to remove the AI CEO. Do they pick a new model? Feed it some new prompts? Alternate between human and AI CEOs?