The term gets used in so many different ways that it has started to mean almost nothing. Team-building workshops invoke it, culture decks cite it, and leaders who have never read Edmondson's original research drop it into conversations as shorthand for a vague aspiration toward niceness.
That's not what it is.
Amy Edmondson at Harvard defined psychological safety as a shared belief that a team is safe for interpersonal risk-taking. Her 1999 research, later reinforced by Google's Project Aristotle, found it was the strongest predictor of team performance, not because it makes teams comfortable, but because it makes them willing to learn. Teams that felt safe to speak up engaged in more learning behaviors: seeking feedback, experimenting, discussing errors, asking for help. Those behaviors drove performance. The connection wasn't direct. It ran through learning. That matters because it tells you what you're actually building when you build psychological safety: not a pleasant environment, but a team that can tell the truth, adjust, and improve.
Edmondson is explicit that safety without accountability produces a comfort zone, not a learning zone. When psychological safety is high but standards are low, people feel safe to express ideas but don't take ownership of outcomes. The goal is both: high safety and high accountability together, which she calls the learning zone. That's where the best work happens.
Why it's harder at the top
Every year I work with executive teams, I ask each CEO to estimate how psychologically safe their team is before I've spoken to any team members. Every single time, the CEO's estimate is higher than what the team actually reports. Not occasionally. Every time.
This is structural. The CEO controls resources, career trajectories, and the daily experience of every person in that room. The power gradient between the CEO and their direct reports is real, and it shapes every conversation whether anyone acknowledges it or not. Edmondson's research notes that psychological safety is harder to build at senior levels precisely because the stakes of speaking up are higher. An executive who contradicts the CEO in a meeting isn't just voicing a different opinion. They're taking a career risk. The team does that math constantly, whether the CEO knows it or not.
There's also a selection effect. Everyone in that room got there by projecting competence under pressure. They've spent careers learning how to manage up, how to look certain when they're not, how to read what the powerful person in the room wants to hear. Asking those people to set aside those skills requires more than an open-door invitation.
What it actually looks like when it exists
The teams I've been in where psychological safety was real were not comfortable rooms. They were some of the most uncomfortable rooms I've sat in, because everything was on the table. Someone would say the budget projection was fantasy. Someone else would name the flaw in the product strategy everyone had been too polite to address. A third person would admit they didn't understand something they were supposed to be expert in. And none of those moments cost them. The relationships came out stronger. The decisions got better.
That's what safety actually buys you. Not harmony. Not the absence of friction. The ability to surface the real thing in the room, with enough relational trust that the person who said it doesn't pay for it afterward.
Edmondson identified three consistent leader behaviors that build it. Visible vulnerability, meaning publicly acknowledging mistakes, uncertainties, and limitations. Framing problems as learning opportunities rather than occasions to assign blame. And actively inviting dissenting views rather than passively tolerating them. The difference between those last two is meaningful. Passive tolerance says dissent is acceptable. Active invitation says dissent is valuable. Teams read that distinction clearly.
The CEO's particular role
Because of the power gradient, the CEO's behavior matters more than anyone else's in the room. A CEO who responds to bad news with curiosity gives the team permission to bring bad news. A CEO who gets visibly defensive when challenged gives the team an instruction: challenge carefully, or not at all.
I've sat with CEOs who were certain their teams felt safe, and then watched the same CEOs, in a group session ten minutes later, visibly tighten when someone pushed back on their read of a situation. The team saw it. The CEO didn't. That gap, between how leaders think they respond and how they actually respond under mild pressure, is where most of the safety work lives.
The CEO who asks for honest input and then debates every dissenting view into submission has taught the team that the input isn't being sought. The confirmation is.
The most reliable test of psychological safety on an executive team isn't a survey. It's this question: when did someone last tell the CEO something the CEO didn't want to hear, and what happened next? The team knows the answer to that question better than the CEO does. And their answer is the operating manual for whether honesty in that room is safe or expensive.