Psychological Safety Is Not Emotional Caretaking

On the limits — and the real substance — of leadership responsibility

Psychological safety has become one of the most frequently cited concepts in leadership culture, and one of the least carefully handled. Originally, it described something concrete: the shared belief that you can speak, question, disagree, or surface problems without fear of retaliation or humiliation. Amy Edmondson, who defined and empirically validated the construct, framed it precisely as the ability to take interpersonal risks at work — to admit a mistake, raise a concern, ask a question — without fear of being punished or demeaned (Edmondson, 1999). It named a real organizational risk: environments where silence becomes adaptive, not because people lack ideas or courage, but because the cost of speaking feels unpredictable. Edmondson's own early research made the stakes vivid — the best hospital teams in her studies reported more errors, not because they made more, but because it was safe enough to surface them. Later work found that the large majority of employees have, at some point, withheld important information from a manager out of fear of the consequences (Morrison & Milliken, 2000).

That framing still matters. Leaders do shape environments, and power dynamics are real. When consequences are opaque or inconsistently applied, people protect themselves. Performance narrows. Trust erodes. Psychological safety gave language to those dynamics, and for many people that language was clarifying rather than indulgent.

When safety becomes emotional caretaking

The trouble is that psychological safety is increasingly treated as a mandate for leaders to manage the emotional experience of everyone around them. Discomfort, disappointment, or insecurity get framed as evidence of leadership failure rather than as an inevitable part of professional life. In that shift, safety quietly becomes caretaking — and this is precisely the distortion Edmondson herself has spent years trying to correct. In a recent analysis of what people get wrong about the concept, she and her colleague describe how the term has been diluted to the point where it can harm the teams it was meant to help, and warn that an over-focus on safety can backfire by pushing teams toward comfort instead of clarity (Edmondson & Kerrissey, 2025).

When leaders are expected to absorb, stabilize, and pre-empt the internal reactions of others, the work of leadership changes. It moves away from setting conditions and toward regulating emotions. This is not a trivial addition to the job; it is a category of labor in its own right. The sociologist Arlie Hochschild named it emotional labor — the continuous, often invisible work of managing feeling to meet the expectations of a role — and documented how depleting it becomes when sustained over time (Hochschild, 1983). Asking leaders to perform that labor on behalf of every person around them is not only unsustainable. It subtly undermines the capacity of adults to engage with complexity, feedback, and disagreement.

This is not an argument for harshness or indifference. It is a boundary. Emotional awareness is a leadership skill; the ongoing emotional management of others is not the same thing, and treating the two as identical does a disservice to both.

What leaders actually owe

Leadership responsibility is real, but it is also specific. Leaders owe their teams clarity, consistency, and fairness. They are responsible for making the rules of engagement explicit and for ensuring that disagreement does not result in retaliation, humiliation, or hidden penalties. They set the tone for how power is exercised and how conflict is handled. There is a strong evidence base for why this is the lever that matters: research on procedural justice shows that people will accept difficult decisions and tolerate unwelcome outcomes when they trust that the process producing them is fair, consistent, and unbiased (Tyler & Lind, 1992). The trust is in the system, not in the pleasantness of any given moment.

What leaders cannot do — nor should they be asked to do — is eliminate emotional friction. Feedback will land imperfectly. Decisions will disappoint. Conflict will activate insecurity. None of this automatically signals an unsafe environment. In fact, Edmondson is explicit that psychologically safe teams have more candid disagreement, not less, and that the experience does not always feel good in the moment; safety enables productive friction rather than removing it (Edmondson, 2019). When leadership is instead defined by the absence of discomfort, organizations lose their capacity for honest exchange and rigorous thinking. Strong leadership creates conditions where people can tolerate challenge precisely because the system itself is trustworthy.

Agency: the missing conversation

There is a second half to this that rarely gets said aloud. If a person's motivation, well-being, or internal stability depends on someone else behaving in a particular way before they can function, agency has already been compromised. This is essentially a question of locus of control — the degree to which people experience outcomes as governed by their own actions versus by external forces. Decades of research link a predominantly external locus of control to lower resilience and diminished initiative (Rotter, 1966). None of this means people should tolerate abuse or genuine instability. It does mean that psychological strength cannot be fully outsourced.

When psychological safety is framed as something leaders must fully provide before individuals can engage, it subtly shifts people into a passive position. Over time, that erodes the very capacities required for meaningful contribution: resilience, discernment, and self-trust. The concept was meant to free people to participate, not to make their participation contingent on someone else first managing their inner state.

What psychological safety is — and isn't

A psychologically safe environment is not one where tension disappears. It is one where the fear of arbitrary harm does. People can engage, disagree, and take risks when they trust that consequences will be fair and proportionate rather than personal or punitive. Confusing safety with comfort flattens this distinction entirely. It turns every emotional response into a diagnostic signal and every leader into a potential emotional caretaker. Edmondson built this very distinction into her framework: she maps teams along two axes — psychological safety and accountability — and shows that high safety paired with low standards yields not high performance but a "comfort zone," while the productive state, the "learning zone," requires high safety and high accountability together (Edmondson, 2019). Strip out the accountability axis and psychological safety collapses into a synonym for "please don't challenge me" — which is exactly how it loses its precision and its power.

The shared responsibility

Psychological safety works when responsibility is shared. Leaders set the conditions under which people work; individuals remain responsible for how they interpret, metabolize, and respond to those conditions. Edmondson herself pairs safety with what she calls psychological ownership — an internal commitment to uphold standards and contribute to the team's goals. When all emotional difficulty is externalized upward, both leadership and followership are weakened: leaders burn out performing impossible emotional labor, and individuals lose the muscle of self-regulation.

At its best, psychological safety protects participation. It allows people to speak without fear while still asking them to engage as adults capable of navigating complexity. That balance is far more empowering than the caretaking model that has grown up around the term — and in the long run it produces healthier leaders, stronger teams, and organizations capable of real work rather than emotional management.

References

Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.

Edmondson, A. C. (2019). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. Wiley.

Edmondson, A. C., & Kerrissey, M. (2025). What people get wrong about psychological safety. Harvard Business Review.

Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. University of California Press.

Morrison, E. W., & Milliken, F. J. (2000). Organizational silence: A barrier to change and development in a pluralistic world. Academy of Management Review, 25(4), 706–725.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs: General and Applied, 80(1), 1–28.

Tyler, T. R., & Lind, E. A. (1992). A relational model of authority in groups. Advances in Experimental Social Psychology, 25, 115–191.