Psychological Safety Has Become a Comfort Blanket
Psychological Safety Has Become a Comfort Blanket, and That’s a Structural Failure, not a Moral One
Psychological safety was never meant to be soft. It came out of research into high-stakes teams – Amy Edmondson’s work on hospital units where staff needed to report errors without fear – and it described something rigorous: the conditions under which people will surface risk, admit mistakes, and challenge a bad call before it becomes a failure.
It was about making the truth sayable, not about making people comfortable.
Somewhere along the way, the concept got hollowed out. The version now circulating in much of corporate life treats safety as a guarantee of comfort, a shield from challenge, a reason to soften standards rather than a mechanism for raising them.
And the usual explanation for this drift, that it was diluted by shallow commentators who never led a real team, is both unfair and, more importantly, wrong.
The dilution wasn’t caused by bad people.
It was selected by the system for the same reason every comfortable misreading of a demanding idea gets selected for.
Why the soft version wins
A demanding concept survives contact with an organisation only if the organisation’s architecture can tolerate it. Real psychological safety is demanding: It requires leaders to hear things they’d rather not, to act on uncomfortable signals, to maintain standards while inviting challenge.
That’s structurally expensive. The comfortable version asks none of it, and that’s precisely why it spreads.
Comfort-based safety is popular for the same reason SWOT and a dozen other diluted tools are popular: it lets an organisation perform a virtue without paying for it.
A leader can declare the team “psychologically safe,” run the engagement survey, and change nothing about who actually gets heard or what happens when someone raises a hard truth. The diluted version is the version that doesn’t threaten the existing architecture, so the architecture keeps it alive.
This isn’t a story about who corrupted the idea. It’s a story about which version of the idea a comfort-seeking system naturally protects.
Which means the fix isn’t to blame anyone for the misreading. It’s to understand why the misreading is so structurally convenient and then to rebuild the concept on the demanding foundation it had to begin with.
What psychological safety actually is
The original idea rests on conditions most “safe” cultures never build:
Clarity of expectations: People feel safe when they know what good actually looks like, not when judgement is withheld. Ambiguity isn’t kindness; it’s the most expensive uncertainty in organisational life.
Competence-based trust: Safety is earned through reliable delivery and reinforced by the structure, not declared in a values statement.
Reciprocity: Everyone shapes the climate, not just the leader. Safety that flows in only one direction is just permission.
Real consequences: Standards hold, performance counts, and behaviour has results. A “safe” environment with no consequences isn’t safe; it’s unanchored.
Productive tension: Disagreement isn’t harm, challenge isn’t aggression, and tough feedback isn’t trauma. The whole point of the original concept was that people could risk friction in service of the truth.
This is the version used by teams where failure has real consequences, e.g. Aviation crews, surgical teams, and elite sports. Not because it feels good, but because it’s the only kind that works when the stakes are real.
What the comfortable version produces
When an organisation adopts safety-as-comfort, the pattern is consistent and quietly corrosive.
People speak up, but nothing changes, because listening replaced deciding.
Standards soften to avoid discomfort.
Accountability gets quietly recoded as cruelty, so it stops happening.
Conflict is avoided rather than used.
The organisation ends up with all the signals of a healthy culture: People sharing, leaders nodding, and none of the function the concept was meant to deliver.
And here’s the part that matters for diagnosis: None of this is a failure of the people in it. Leaders adopt the soft version because they were told it was the responsible thing, and because the architecture around them rewards the appearance of care over the substance of it. The teams emote rather than execute because the system gave them permission to share without ever requiring them to be accountable for what they share. The behaviour is, as always, a rational response to what the structure actually rewards.
Rebuilding it: safety anchored to accountability
If psychological safety is going to be useful again, it has to be rebuilt around respect and accountability rather than comfort. The reframe is simple to state and demanding to live:
You are safe to speak and are responsible for the quality of what you contribute.
You are safe to challenge and are expected to be challenge-ready yourself.
You are safe to take risks, and you own the outcomes.
You are safe as a person and are required to behave as an adult, with the respect that implies, in both directions.
That’s not a softening of the original concept. It’s a restoration of it. Edmondson’s safe teams weren’t comfortable; they were candid, and candour requires exactly this pairing: The freedom to speak and the obligation to be worth listening to.
What leaders actually have to do
Reclaiming psychological safety isn’t a rejection of care, empathy, or humanity.
It’s a rejection of the hollow version that has been mistaken for them.
The work is to define safety so that it strengthens performance rather than excusing its absence: clarity that makes expectations knowable, consequences that make standards real, reciprocity that makes the climate everyone’s responsibility, and enough genuine respect that challenge can happen without being mistaken for attack.
But the deeper point is the one the comfort-versus-rigour debate usually misses.
You cannot exhort your way to real psychological safety any more than you can exhort your way to candour, accountability, or trust.
Telling leaders to “be more rigorous about safety” fails for the same reason every behavioural instruction fails against a contrary architecture.
Real safety appears when the structure makes truth-telling survivable: When raising a hard signal doesn’t end a career, when the person who challenges the plan isn’t quietly sidelined, when accountability is built into the system rather than performed by a brave individual.
Safety isn’t a posture leaders adopt. It’s a property of the architecture they build.
Respect people. Hold them accountable. And design the system so that the two reinforce each other rather than trading off.
That’s the psychological safety worth having and it was the original version all along.