Shadow AI Isn’t a Tech Problem.
It's a Leadership Problem
This might actually make some of you angry, but I think it’s important enough to risk it.
I’m a country boy from South Dakota, so my default is to keep things simple and sometimes a little too blunt.
Too many people are letting AI do the thinking they should still be doing themselves.
And the most concerning part? It’s happening in the shadows.
That’s the nature of shadow AI—it doesn’t show up in your governance frameworks, your risk reports, or your board decks. For writers, it can be a powerful tool. It can quietly shape drafts, headlines, and research summaries, and it can even influence strategic messaging before anyone stops to ask who actually did the thinking. Shadow AI is there, and it is influencing decisions that matter.
Let me be clear: I’m not anti-AI. Not even close.
I believe AI is one of the most powerful tools we’ve ever had. I actively encourage teams, executives, and organizations to use it. Use it and get as much horsepower from it as you can.
But there’s a difference between using AI and outsourcing judgment to it.
Right now, shadow AI is increasingly being used to:
Influence forecasts.
Build sales pipelines.
Draft strategic recommendations.
Shape executive decisions.
And way too often, there isn’t enough human review, challenge, or accountability.
In some cases, leadership doesn’t even know it’s happening.
That’s where the dangerous implications start—especially in regulated industries like banking and credit unions.
Here’s the question I keep coming back to:
Are you ready to bet your career on what an AI is telling you to tell your board or your boss?
Because that’s what this comes down to.
If the model is wrong, biased, incomplete, or just wildly and confidently off base, the AI doesn’t take the hit.
You do.
And when organizations start trusting AI outputs without enough human engagement, the implications build quickly:
Forecasts get built on untested assumptions.
Pipeline decisions prioritize the wrong things.
Leadership accepts polished answers instead of the right ones.
Teams stop asking hard questions.
At that point, in my view, we’re not talking about efficiency anymore.
We’re talking about abdication. And there are a lot of organizations out there that have already abdicated and don’t even know it.
In regulated environments, this is a leadership failure waiting to happen.
Think about it:
Who is accountable for AI-influenced decisions?
Can you evidence how a recommendation was formed?
Do you know when AI was used—and when it wasn’t?
Could you defend those decisions to an examiner, an auditor, or in a courtroom?
If the answer is unclear, you more than likely have a problem.
And regulators, courtrooms, or boardrooms are not going to accept “the AI suggested it” as an answer.
AI should support human decision-making—not replace the people responsible for the decision.
That means:
Clear accountability for outcomes.
Visibility into where AI is used.
The ability to challenge and verify outputs.
A culture that rewards questioning, not blind acceptance.
In my company, somebody will always be accountable.
That’s the rule.
Somebody has to own the decision. That’s not optional.
If you’re at a regulated organization, here’s a practical place to start:
I’m offering a 30-minute Shadow AI Risk Review for a limited number of teams.
We’ll walk through:
Where shadow AI is most likely already showing up in your organization.
The decision points most at risk, including forecasting, lending, marketing, and operations.
Gaps in accountability, visibility, and auditability.
Just a direct conversation about where you stand and what to do next.
If you’re interested, reach out directly here:
Because at the end of the day, this isn’t about stopping AI. AI is here to stay.
The challenge is whether organizations will use it with discipline, transparency, and accountability—or let it quietly take over the thinking they should still own.