Psychological safety has become a phrase leaders say and employees test.

The saying happens in all-hands meetings and values decks. The testing happens somewhere else entirely — in small moments no one announces: what happened to the last person who brought bad news, whether the quiet coworker got curiosity or relief, which logos came back in June and which quietly didn't.

I've spent more than twenty-five years in leadership — retail, banking, HR, people strategy — and most of what I've learned about safety I learned by watching the distance between the phrase and the feeling. This page is about that distance: what psychological safety actually looks like, how it erodes without anyone deciding to erode it, and what rebuilding it asks of anyone with standing in a room.

What Psychological Safety Is — and Isn't

The term comes from Harvard's Amy Edmondson, who defined team psychological safety as a shared belief that the team is safe for interpersonal risk-taking. In plain language: people believe that speaking up — with a question, a mistake, a concern, an idea, a fact about who they are — will not cost them. Google's Project Aristotle famously studied what made its best teams work and landed on the same answer: psychological safety mattered more than who was on the team.

Two things it isn't. It isn't comfort. Safe teams argue more, not less — they challenge each other harder because disagreement isn't dangerous. And it isn't a policy. You cannot announce psychological safety, because it isn't a rule; it's a belief. Beliefs are built from evidence, and the people around you have been collecting evidence the whole time.

Which means the real question isn't whether you've said the words. It's what evidence you've been producing.

Why Safety Sits at the Center of This Project

The Human Leader Project™ runs on a single premise: your impact is measured by what the people you affect actually experience — not by what you intended, announced, or achieved. Psychological safety is where that premise gets tested first, because safety is the precondition for every other human experience I write about. Nobody feels seen in a room where being seen is dangerous. Nobody belongs in a room where belonging has a price. And dignity — the conviction underneath all of this work — can't survive where people have to perform their way into being treated as whole.

Safety also decides something more personal to this project: which story gets told where. In an unsafe room, the honest version of your leadership doesn't disappear — it relocates. The meeting stays polite, and the truth moves to the dinner table, where you will never hear it. The less safety you build, the wider the gap between the story people tell you and the story they tell about you. Every signal in the next section is, in the end, a person deciding which of those two stories you've earned.

The Signals People Actually Test

Nobody runs their safety assessment in a survey. They count signals — small, specific, and mostly invisible to the person being measured. These are the ones I've watched people count, and counted myself:

  • What happens to the messenger. The first person to bring bad news gets a verdict, and everyone watching remembers it longer than you do. One flash of blame buys months of filtered information.

  • Whether silence gets curiosity or relief. When someone stops raising concerns, unsafe teams read it as resolution. What actually happened is that the person got trained. The invisible workplace is built in exactly those moments.

  • Whether being invited means being heard. A seat at the table is a fact; belonging is what happens when the person speaks. Watch what the room does with their voice — everyone else is.

  • Whether the symbols come back. I once went looking for two Pride logos that hadn't returned in June and found a quieter grief underneath — the math people do about whether it's safe to be seen, without ever saying a word. Symbols are evidence too. So is their absence.

  • The price of being known. In some rooms, being fully yourself is free. In others it costs, and everyone paying has done the arithmetic. People notice who sets that price.

  • What results excuse. If a leader who produces numbers is allowed to produce fear alongside them, every person watching learns the organization's real values — whatever the values deck says.

How Safety Erodes

Almost never through an event. Through training.

Someone raises a concern once, maybe twice, and nothing happens — so they stop. Someone brings a hard truth and watches it get received as disloyalty — so the next truth stays home. Someone shares a piece of who they are and the room goes slightly, politely cold — so the careful version comes to work from then on. None of these moments look like anything from the outside. The metrics hold. The meetings stay pleasant. The record shows stability, right up until the resignation letter, the union petition, or the exit interview that says nothing at all.

That's the cruel mechanics of it: safety erodes silently because it's eroding. The less safe people feel, the less signal they give you about it.

What Rebuilding Looks Like

Safety is lost in increments, and it's rebuilt the same way — not with an announcement, but with evidence produced consistently enough that beliefs update. Three forms of evidence matter most:

  • Pass the next test in public. The next time bad news reaches you, your reaction is the culture. Thank the messenger specifically, in front of whoever is watching, and deal with the news separately from the person who carried it.

  • Close the loop out loud. "You raised this. Here's what happened." Or just as powerfully: "Here's why we couldn't." Being heard doesn't require getting your way — it requires evidence that speaking reached something.

  • Keep the signals up when it's inconvenient. The symbol that stays through the hard season is worth ten that appear in the easy one. Consistency is the whole message.

And then: repetition. One good response to bad news isn't safety — it's an anomaly until you repeat it. Beliefs built from years of evidence take more than a week of counter-evidence to move. That isn't unfair. It's how trust works everywhere else in life, too.

Read the Essays

The essays behind this page, in the order I'd read them:

One Essay, Every Week

Safety is one thread of the larger question this whole project asks — what the people you affect experience, and what story they tell about it at the dinner table. The Human Leader Project is one honest essay every week, free, with no leadership theater.

The story your people tell when you're not in the room is your truest legacy.

Matthew Barbour