We spend a lot of time worrying about whether the team will adopt AI. We should spend more time worrying about how the executives are already using it.
The biggest AI risk in most businesses is not a junior analyst pasting a spreadsheet into a chatbot. It is a senior leader, driven by curiosity and the urge to look innovative, reaching for whatever unapproved tool is in the news that week and feeding it real company information. No understanding of where the data goes. No oversight. No second thought.
It is an easy trap to fall into. Leadership is meant to be ahead of the curve, and AI makes everyone feel ten steps ahead after one good answer. But confidence is not the same as control, and a clever output is not the same as a safe one.
The answer is not to ban the tools. That only pushes the behaviour underground. The answer is context and guardrails:
- Give people approved tools that are grounded in your own data and governed inside your own tenant.
- Be clear about what may and may not go into a public model.
- Treat AI literacy as a leadership skill, not a junior one.
Shadow AI does not start at the bottom of the org chart. It starts wherever enthusiasm runs ahead of understanding, and that is just as often at the top. Lead by example, or you will become the example.
