
In April 1961, a small group of very capable advisers sat with President John F. Kennedy and approved a plan to send Cuban exiles ashore at the Bay of Pigs. Several of them had doubts. They kept those doubts to themselves. The meetings were cordial, the president was well liked, and the invasion collapsed within days.
Eighteen months later, during the Cuban Missile Crisis, Kennedy ran his advisers through a very different process, one designed to pull disagreement out of people instead of waiting for it. Harvard’s Amy Edmondson and Michaela Kerrissey use the contrast to make a point most organizations still miss. The first group was comfortable. The second group was safe enough to argue.1
Psychological safety has become one of the most cited ideas in management, and one of the most misread. In a lot of companies it now means pleasant meetings, gentle feedback, and an engagement score that trends up. That version feels good and produces very little. The real thing is a shared permission for candor, and it only pays off when it sits next to a high standard. Leaders who want the payoff have to stop asking how comfortable people feel and start counting what people are willing to say.
How a Research Finding Turned Into a Mood
Google’s Project Aristotle studied 180 teams, 115 engineering teams and 65 sales groups, and found that psychological safety mattered more than any other team dynamic it measured, ahead of dependability, structure and clarity, meaning, and impact. In Google’s summary, who was on a team mattered less than how the team worked together.2 The finding traveled fast. It showed up in leadership programs, in engagement surveys, on offsite agendas, and eventually in proposed legislation.
Somewhere along the way the definition went soft. The original idea is plain enough: people believe they can raise ideas, questions, concerns and mistakes without being humiliated or punished for it. In practice that has drifted toward something closer to “people feel good here.” Edmondson and Kerrissey list six common misreadings in their 2025 Harvard Business Review article.1 Two of them do most of the damage inside leadership teams. Safety gets confused with niceness, and safety gets treated as the opposite of accountability.
A Nice Team Can Be a Silent Team
When the norm is to keep things pleasant, disagreement turns into a social cost, and people stop paying it. The meeting goes well. The concern goes home with the person who had it.
Most people who have worked on a large program have sat in this meeting. Every workstream reports green. The sponsor thanks everyone for the progress. Two weeks later the go-live date moves by a quarter, and it turns out four people in that room saw it coming. None of them wanted to be the one who spoiled a good review.
Edmondson made the distinction clear years ago. Psychological safety, she wrote, does not mean people agree with one another for the sake of being nice, and it isn’t about becoming comfortable at work.3 In the 2025 article she and Kerrissey describe it as a shared sense of permission for candor.4 They also write that where it exists, people believe sharing hard truths is expected.5 Expected is the word that matters. A team where hard truths are tolerated if they’re delivered gently is a long way from a team where holding one back counts as letting everyone down.
Comfort Has a Ceiling
There’s research suggesting the comfortable version doesn’t just fail to help. Past a certain point it can hurt. Liat Eldor and Michal Hodor of Tel Aviv University and Peter Cappelli of Wharton analyzed hundreds of internal and external performance ratings across different businesses and jobs over time. Moderate psychological safety improved performance. Beyond a certain level the returns fell away, and productivity and accuracy suffered. Cappelli said he was surprised that people assumed there was no limit to it.6
That fits Edmondson’s own description. Safety, she wrote, takes the brakes off. It is not the fuel that powers the car, and leaders still have to set high standards.3 She and Kerrissey treat safety and accountability as two separate dimensions instead of two ends of one scale, and they’re blunt that high safety paired with low standards is not excellence.1 A team in that corner gets a pleasant place to work and ordinary results.
Leaders Are Holding Back Too
The usual assumption is that safety is something employees lack and leaders hand out. Recent data says leaders feel the pressure as well. In a January 2026 Harris Poll survey for Turas Leadership Consulting, 317 directors and above were asked about seeking input on big decisions. Ninety percent said they wish their teams would more often bring feedback that challenges them or the status quo. Nearly two-thirds said they would ask for input more often if it didn’t risk making them look weak.7
So a lot of the people who want the challenge are holding back the very move that would invite it. The effect shows up further down. In a separate Harris Poll survey of 2,206 employed Americans, reported by David Grossman of The Grossman Group, only 19 percent of people working for leaders they rated as good said they feel heard.8
A leader who doesn’t ask and a team that doesn’t offer end up in the same quiet room. Both sides read the quiet as harmony.
Why Policies and Scores Miss It
In 2024, Rhode Island legislators introduced a Workplace Psychological Safety Act, which stalled in the state senate. Edmondson and Kerrissey use it as an example of a common mistake, the belief that a policy can create safety.1 Their position is that psychological safety is built in a group, interaction by interaction.5 Their research also finds that it varies from team to team inside the same company, even one with a strong overall culture.1
That’s why a company-wide score can mislead. An average can look healthy while a handful of teams sit on exactly the information the business most needs. Dräger’s 2026 Safety and Health at Work report shows how wide the gap between belief and behavior can get. Eighty percent of respondents said their organization manages workplace safety risks effectively, and 73 percent said it runs open communication channels on health and safety. Yet 53 percent said the current economic climate makes them less willing to raise health and safety issues or challenge leadership.9 An open channel that people are nervous about using doesn’t count for much.
What Candor Looks Like in Practice
None of this calls for a new program. It calls for leaders to change what happens in the meetings they already run.
Talk about the work, not the feeling. Edmondson and Kerrissey suggest talking less about psychological safety itself and more about the goal and the context.5 People speak up when they can see why their input matters to an outcome they care about. A leader who opens with “this launch has two ways to fail and I can only see one of them” gets a very different meeting than one who opens with “I want everyone to feel safe here.”
Ask for the disagreement on purpose. Before any significant decision, ask who sees it differently, then wait through the silence that follows. Kennedy’s second process worked because disagreement was part of its design.1 One retail operations executive in the HBR research improved his team’s candor by slowing his own meetings down.1
Watch the response more than the invitation. Every leader says the door is open. Teams learn from what happens to the first person who walks through it. In Edmondson’s informal surveys, managers said only 1 to 4 percent of failures really deserve blame, while 70 to 90 percent get treated as if they were.3 That gap teaches people more than any town hall.
Hold the standard in the same conversation. Safety with a soft standard drifts toward comfort, which is where the Wharton findings suggest performance starts to slip. Pair every invitation to speak up with a clear expectation about the quality of the work, and say both out loud.
Safe Enough to Be Uncomfortable
Psychological safety is still one of the most useful ideas in management. The trouble is the version most organizations have built. It rewards pleasant meetings and high survey scores, and it leaves the hard information where it started, with the person who was nervous about sharing it.
The test is simple and a little uncomfortable. At the next major decision, count how many real objections were raised before it was made, and who raised them. If the count is zero, that room has the same problem Kennedy’s advisers had in the spring of 1961, and it will probably find out the same way.
References
- Amy C. Edmondson and Michaela J. Kerrissey, “What People Get Wrong About Psychological Safety,” Harvard Business Review, 103(3), May-June 2025, pp. 52-59, as summarized in The Field Guide to Psychological Safety research, Psych Safety.
- Google re:Work, “Guide: Understand Team Effectiveness.”
- Amy C. Edmondson, “How Fearless Organizations Succeed,” strategy+business, November 14, 2018, adapted from The Fearless Organization (Wiley, 2018).
- Amy Bernstein, “The Truth About Psychological Safety,” The Leader’s Agenda, Harvard Business Impact, May 21, 2025.
- Adie Fein, “Debunking misconceptions about workplace psychological safety,” Harvard T.H. Chan School of Public Health, June 26, 2025.
- Knowledge at Wharton, “The Downside of Psychological Safety in the Workplace,” November 21, 2023, on Liat Eldor, Michal Hodor, and Peter Cappelli, “The Limits of Psychological Safety: Nonlinear Relationships with Performance.”
- Laurel Kalser, report on leaders who want team input but fear asking for it, HR Dive, February 5, 2026, on a Harris Poll survey for Turas Leadership Consulting.
- David Grossman, “The 19% Problem: Why Good Leaders Think They’re Listening When Employees Disagree,” Allwork.Space, May 14, 2026.
- Adam Pope, article on psychological safety and the intention gap, British Safety Council, July 30, 2026, on the Dräger Safety and Health at Work 2026 report.








