Principal Resolution

Somewhere around 2018, psychological safety became the kind of phrase that gets written on a whiteboard in a sans-serif font and never quite makes it back off the wall. This is not an argument against the concept. It is an argument against what we have done with it.

Psychological Safety Walked Into a Meeting

Somewhere around 2018, "psychological safety" became the kind of phrase that gets written on a whiteboard in a sans-serif font and never quite makes it back off the wall. If you work in any organization of appreciable size, you have been in a room where someone used it. You may have used it yourself. You nodded. The person across the table nodded. The consultant facilitating the offsite wrote it next to "vulnerability-based trust" and drew an arrow to a circle labeled "high performance."

Nobody said the thing they had been not saying for eighteen months.

This is not an argument against psychological safety. It is one of the most well-researched, practically important concepts in organizational behavior, and the evidence for its impact on team performance is robust and consistent. This is an argument against what we have done with it — which is to take a genuinely useful idea, dress it in the vocabulary of corporate transformation, and deploy it in exactly the environments where it cannot survive.

What It Actually Is

Amy Edmondson's original research defined psychological safety as the belief that one will not be punished or humiliated for speaking up with ideas, questions, concerns, or mistakes. That is a precise definition. It is about a specific kind of risk — the interpersonal risk of saying something that might make you look bad — and about whether people in a particular environment believe that risk is present or absent.

It is not a mood. It is not a vibe. It is not the feeling that everything is fine and everyone likes each other. Some of the most psychologically safe teams on record are also the most combative — they argue constantly, they challenge each other's assumptions relentlessly, they push back on leaders without apparent hesitation. They do this because they believe the relationship is strong enough to hold it. They have tested that belief and it has held.

The organizations that have misunderstood this — and there are many — have produced environments that feel psychologically safe because no one says anything difficult. The meetings are cordial. The feedback is gentle. The retrospectives are upbeat. And the problems that needed to be named in Q3 are now Q1 crises.

Comfort is not safety. Conflict avoidance is not safety. The absence of friction is not safety. These are, in many cases, the precise opposite of safety — because they are what happens when people have concluded that the risk of speaking is too high, and have simply stopped trying.

A Brief Tour of What This Looks Like Across Industries

Manufacturing

The floor supervisor at a mid-sized fabrication plant has been watching the same quality defect pattern for six weeks. He knows what is causing it. He also knows that the last person who raised a process concern to the plant manager did not get fired — nothing that dramatic — but the relationship shifted in a way that was unmistakable to everyone watching. The monthly one-on-ones got shorter. The tone got colder. Nothing was ever said explicitly.

So the supervisor tracks the defect. He logs it. He builds a mental case that is watertight. And he waits for an environment in which it is safe to surface it, which never quite arrives. The defect becomes a recall. The recall becomes a news story. The plant manager, in the subsequent investigation, says he had no idea.

He did not have no idea. He had a culture.

Corporate

The product manager has been in enough sprint retrospectives to know how they go. The facilitator asks what went well. People say things. The facilitator asks what could be improved. People say slightly milder versions of the same things they said last time. The action items from the last retrospective, which were never acted on, are quietly dropped from the agenda.

No one says: we have been shipping features that do not solve the actual problem because the roadmap is being driven by a stakeholder relationship rather than user evidence, and everyone in this room knows it, and none of us have said it directly because the last person who pushed back on that stakeholder got quietly moved off the team.

The retrospective ends. Everyone files out. The next sprint starts. The pattern continues.

Nonprofit

The program director has run the numbers three times because she cannot quite believe them. The flagship initiative — the one that anchors the organization's fundraising narrative, the one the executive director has been presenting to donors for four years — is not producing the outcomes it was designed to produce. The data is clear. The methodology is sound. She knows what she found.

She also knows that this program is the organization's identity. She knows the executive director's name is professionally tied to it. She knows that the last person who brought genuinely critical program evaluation to the board meeting left the organization within eight months — not fired, just gradually, and everyone understood why.

So she writes a careful report. She highlights the successes, which are real but partial. She describes the challenges, which are also real, in language that preserves optionality. She recommends further study. The board receives it warmly. The donors are not told.

The program continues. The outcomes do not improve. Eventually a journalist finds the data.

Healthcare

The circulating nurse has been in enough operating rooms to notice patterns. She has noticed that one of the surgeons — highly regarded, technically excellent, genuinely difficult — has a communication style in the OR that causes the rest of the team to go quiet at exactly the moments when someone should be speaking up. She has watched nurses hesitate to call a count discrepancy because of how he responded the last time. She has watched residents stay silent rather than flag a concern because the cost of being wrong in front of him is too high to risk.

She has mentioned it. Once, to a charge nurse, who said she understood and would look into it. That was fourteen months ago.

The protocol exists. The checklist exists. The culture does not match them. In the specific, charged environment of that OR, with that surgeon, the checklist is a formality and everyone knows it.

This is the version of psychological safety that has a body count. It is not a metaphor.

The HR-ism Problem

Here is what happened to psychological safety in most organizations: it got institutionalized before it got understood.

Someone read the research — or a summary of the research, or a summary of a summary — and concluded that psychological safety was a thing organizations needed to have. They brought it into the organization through the legitimate channels: the L&D function, the leadership development curriculum, the engagement survey. They created training programs and facilitated workshops and added it to the competency model.

And then a strange thing happened. In the organizations that did this, psychological safety became a performance criterion. Leaders were assessed on whether they were creating it. Teams were surveyed on whether they experienced it. Scores were reported. Action plans were developed.

The result, in many cases, was that leaders became very skilled at performing the behaviors associated with psychological safety — active listening, validating language, inviting dissent in the abstract — while remaining fundamentally unchanged in how they actually responded when someone said something they did not want to hear.

This is the gap that the training cannot close, because the training addresses behavior and the problem is belief. If a leader believes, at some level, that dissent is a problem to be managed rather than information to be processed, no amount of skill-building will produce a psychologically safe environment. The team is not reading the leader's technique. They are reading the leader's response. They are asking: what happens to people who say the true thing here.

What Actually Creates It

The research on this is fairly consistent: psychological safety is built through demonstrated response, not declared intent.

When a leader responds to a mistake by focusing on what can be learned rather than who is to blame, they are building it. When a leader responds to a challenge to their position by engaging with the substance rather than the standing of the person who raised it, they are building it. When a leader responds to bad news by thanking the person who delivered it — visibly, in front of others — they are building it.

When a leader responds to any of these things by punishing, dismissing, marginalizing, or subtly cooling the relationship, they are destroying it. And because the team is watching, they are destroying it not just with the person who experienced it, but with everyone who observed it.

This is why psychological safety is a leadership problem, not a culture problem. Culture is downstream of what leaders do, not the other way around. You do not get to a high-trust, high-candor culture and then find that leaders can operate that way in it. You get leaders who operate that way, and the culture assembles around them.

The frameworks and workshops are useful for naming the concept and giving people vocabulary. They are not capable of producing the thing itself. The thing itself is produced by what happens in specific moments, with specific people, when the truth is inconvenient.

The Practical Question

If you are a leader reading this and wondering whether your team experiences psychological safety — not as a score on a survey, but as a living reality — there is one question worth sitting with:

When did someone last tell you something you did not want to hear?

Not in a structured feedback session. Not in a 360 debrief. In a real conversation, about something real, that required them to take a genuine interpersonal risk to surface it.

If you cannot remember, there are two possible explanations. One is that everything in your environment is actually fine and there is nothing difficult to say. The other is that the people around you have concluded that saying it is not worth the risk.

One of those explanations is far more common than the other.

The question is not whether your organization talks about psychological safety. It is whether the people in it believe — based on evidence, not declarations — that the cost of telling the truth is lower than the cost of staying quiet.

If they do not believe that, the whiteboard does not help.