The Trust Gap

Every conversation about AI seems to end up in the same place: legal.

Can we use this image?

Can we publish this?

Can we train on that dataset?

What are the disclosure requirements?

They're important questions, and every company should be asking them. But I've started to wonder if they're also giving us a false sense of security. Just because something is legally defensible doesn't automatically make it a good decision.

"Can we?" has never been the question that built great brands.

Somewhere along the way, we've turned AI ethics into a compliance exercise. A checklist. A policy document. A meeting that happens right before launch.

I think that's backwards.

Ethics isn't where a project ends. It's where leadership begins.

Lawyers protect organizations from unnecessary risk. That's their job, and they're incredibly good at it. Leaders have a different responsibility. They decide what kind of company they're building and what kind of relationship they want to have with customers. Those aren't legal questions. They're cultural ones.

The irony is that we rarely talk about AI ethics when things are going well. We only notice it after something breaks—a synthetic CEO voice that crosses a line, a customer who didn't realize they were chatting with AI, a campaign trained on work that was never meant to become training data, or another headline about copyright.

By then, ethics has already become damage control.

It was supposed to be product strategy.

James Clear once wrote that every action you take is a vote for the type of person you wish to become. Brands work the same way. Every AI decision, no matter how small, is a vote for the kind of company you're becoming.

Most of those votes don't feel significant in the moment. One automated email. One AI-generated product image. One chatbot response that sounds just a little too robotic. One disclosure you quietly decide isn't necessary because no one will probably notice.

Individually, they're forgettable.

Collectively, they become culture.

I don't think AI is going to destroy trust overnight. That's too dramatic.

What's more likely is something far less exciting and much more dangerous: a slow erosion of authenticity through a thousand tiny shortcuts.

The email that doesn't quite sound like your brand.

The customer service interaction that optimizes for speed instead of empathy.

The campaign that's perfectly polished but somehow forgettable.

None of those moments become front-page news. They simply make people feel a little less connected to you than they did yesterday.

Trust rarely disappears all at once.

It wears away quietly.

The companies that stand out over the next decade won't necessarily be the ones using the most AI. That's a race with no finish line. The models will improve, competitors will catch up, and today's breakthrough will become tomorrow's default setting.

The real competitive advantage will be judgment.

Knowing when automation genuinely improves an experience and when it quietly diminishes one. Knowing when efficiency serves the customer—and when it only serves the quarterly report. Knowing when a human voice is worth protecting, even if it's slower and more expensive.

Those aren't technology decisions.

They're leadership decisions.

Lately, I've started asking a different question whenever AI enters a project.

Not, "Can we do this?"

Instead: "If our customer watched exactly how this was made, would they trust us more or less?"

I've found the answer usually arrives before the debate does.

That little feeling in your stomach is often more valuable than another policy document.

AI isn't forcing us to rethink technology.

It's forcing us to clarify who we are.

The models will keep getting smarter.

The question is whether our judgment keeps pace.