Plan for "AI goes bad" as an explicit scenario.

Widespread use of AI may well infect organization's operations and decision making processes with increased uncertainty and legal liability.

Also, the risks associated with simply gifting vendors so much influence and control should be obvious.

Example: What happens if (or rather when imo) vendors decide to dramatically increase their rates?