Psychological safety emerged as a concept for organisational effectiveness in the 1990s, when Harvard Business School Professor Amy Edmondson coined the term “team psychological safety”. It gained prominence in 2012, when Google's Project Aristotle identified it as the key factor that distinguished high-performing teams from those with lower performance. The concept described workplace environments where people felt able to raise concerns, ask questions, admit mistakes, and propose ideas without fear of negative consequences.

UNSW Business School’s Professor Karin Sanders recently spoke with Prof. Edmondson about psychological safety in contemporary organisations. They discussed how uncertainty affects workplace dynamics, the particular challenges of remote and distributed work, common misconceptions that hinder organisational progress, the team-based nature of psychological safety, and the implications of artificial intelligence for team environments. The conversation offered valuable insights for leaders seeking to create environments where people can speak up, share ideas, and work through uncertainty together.

Prof. Sanders: Given that we are in a more unpredictable business climate than in previous decades, how has psychological safety evolved? Is it more difficult to achieve now than in more stable times?

Prof. Edmondson: The research shows that psychological safety is even more important when uncertainty is greater. The greater the uncertainty, the more knowledge-intensive and complex the work is, and the larger the effect of psychological safety on performance. It matters more because there are more things that could go wrong in unpredictable environments, so organisations need people speaking up, generating ideas, and experimenting to navigate through ambiguity.

This creates a paradox that leaders must navigate. Organisations need psychological safety more in uncertain times, yet it can be harder to build precisely because unpredictable conditions are anxiety-provoking. The solution to working through this paradox is to make it discussable. The act of naming the challenge and the risk that organisations face puts everyone on the same page. It is a form of shared acknowledgement. Teams do not have to feel positive about the entire world or external conditions, but they can feel positive about each other if they start laying things out and begin to address challenges honestly with one another about what they are up against.

Learn more: How to manage psychosocial hazards in the workplace

Prof. Sanders: You mentioned naming the challenges as a starting point. What should leaders do next after they have named what teams face?

Prof. Edmondson: Leaders cannot approach this challenge with the idea that they can fix it or create psychological safety on their own. If leaders believe that creating psychological safety is their responsibility alone, they will fail. That approach puts too much burden on one person and misunderstands the nature of psychological safety. Leaders need to open up the challenge and make it a team sport, inviting others in to help solve the problem and build the best possible climate together.

This collaborative approach recognises that psychological safety emerges from group dynamics and interaction patterns, not from individual mandates or policies handed down from above. Leaders set the stage and create conditions, but the team collectively builds the environment through how members treat each other, respond to ideas, and handle disagreements and errors.

Remote work and team dynamics

Prof. Sanders: Is building psychological safety more difficult when people work remotely and are in the office on different days, without consistent face-to-face contact?

Prof. Edmondson: There is no question about this. The data is clear that distributed or remote work inhibits candour and inhibits psychological safety. If you think of psychological safety as ease of speaking up – and I do think of it that way – even mechanically, ease of speaking up is lower in remote formats. When teams can be together face-to-face, they should be together. When they cannot, structure becomes essential.

Leaders have to almost go overboard on structure in remote settings. This might mean going around the group systematically to hear from everyone, being deliberate and even heavy-handed in asking questions, and using people's names to make the cost of speaking up lower and the cost of silence higher. In remote settings, you must be more of a process manager, rather than letting conversations be organic and emergent, as they can be when people are face-to-face. Leaders must use available tools like chat functions deliberately and recognise that structure becomes your friend in distributed work environments.

Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School

Prof. Sanders: Given these challenges with remote work, does this mean that, for psychological safety, it is better that people work more in the office and less remotely?

Prof. Edmondson: Not necessarily, because psychological safety is not the ultimate goal in itself. Psychological safety is a means to an end. The end is performance, innovation, learning, and adaptation. Face-to-face work is more important for certain kinds of work, particularly ideation, conceptualising, and planning in an open space where ideas are still forming. When there is a need to diverge before converging – to generate many possibilities before selecting among them – that work is substantially better done face-to-face.

For a quick check-in meeting, remote formats work adequately. For one-on-one conversations with fairly routine content, remote formats are sufficient, especially if what is being discussed is familiar territory. However, the less routine the work, the more ideation is required, the more conceptualising is involved, the more structuring something from nothing is necessary, and the more being together face-to-face will make a difference in the quality of the output. The mode of work should match the nature of the task, and organisations should be thoughtful about which work requires which setting.

Misconceptions about psychological safety

Prof. Sanders: There are various misconceptions about psychological safety circulating in organisations. Do these misconceptions hinder organisational progress?

Prof. Edmondson: Misconceptions absolutely matter because they have an impact. One common misconception is that psychological safety means everybody should feel comfortable all the time and everyone should just be nice. If leaders hold this misconception, they will not consider it proper or effective to establish psychological safety, because clearly, much work is uncomfortable, feedback is uncomfortable, and disagreement is also uncomfortable. Yet feedback, disagreement, and sustained effort through difficult challenges all lead to excellence. Organisations need these elements.

When leaders hold this misconception, they feel a contradiction. They know they need to push for results and have difficult conversations, but they think psychological safety means avoiding discomfort. The way many leaders resolve that contradiction is to pretend they are implementing psychological safety, but not really commit to it, because it does not make sense to them, given their understanding. However, if leaders think of psychological safety differently – as a learning environment, one where people feel genuine permission for respectful candour – then they can see exactly why that might be important and why it might lead to better outcomes. Then they become more interested and willing to do the hard work it takes to bring it about. These misconceptions can literally get in the way of people's willingness to do what it takes to establish psychological safety effectively.

Learn more: How mindfulness can help tackle healthcare’s burnout crisis

Prof. Sanders: Can you describe another common misconception about psychological safety that you encounter?

Prof. Edmondson: Another misconception that comes up frequently occurs when people say to their leaders that, because the leader did not agree with them in a meeting, the leader took away their psychological safety. This reflects a misunderstanding that psychological safety means whatever someone says will prevail, which is simply not true or realistic. Agreement on outcomes is not a measure of psychological safety.

If a leader disagrees with someone, they are not taking away psychological safety. Rather, they are demonstrating that psychological safety exists, because disagreeing is not easy. Giving feedback is not easy. We are more likely to disagree and provide feedback when we believe it is expected, welcome, and necessary, even though it remains uncomfortable. This distinction is critical for organisations to understand. Psychological safety creates permission for disagreement, not protection from it.

The team-based nature of psychological safety

Prof. Sanders: What organisations in the US or Australia are doing an excellent job of showing the importance of psychological safety and demonstrating that it can work without misconceptions?

Prof. Edmondson: The challenge with naming organisations is that psychological safety is always variegated, meaning there are always differences across groups within any organisation. Research has shown that psychological safety varies substantially within organisations. Except in small start-ups with perhaps a dozen people, organisations do not have consistent high psychological safety across all groups. There are always areas that perform better and areas that perform worse.

This is not just a research data point or a limitation of measurement. It is actually a concept that really matters. Psychological safety is team-based, not organisation-wide. Think about an organisation like Microsoft or Google. They are vast enterprises, with thousands of teams and hundreds of thousands of employees. Psychological safety describes the emergent climate in my team and your team – specific, local environments where people work together regularly.

Learn more: Psychosocial safety: Addressing hazards and enhancing wellbeing

From a management perspective, this means that all a leader really needs to do is focus on their own team. Leaders cannot fix the universe, but they can help their team. They can set the stage, solicit input, develop the skills of high-quality conversations, and make progress on what will inevitably be a bumpy road ahead. If each leader does that work, each team member is aligned, and everyone else is doing that throughout the organisation, they create cultures that are capable of pursuing excellence.

Psychological safety is built from the ground up – it is not a top-down mandate. Leaders at the top cannot fix the universe of team dynamics throughout a large organisation. But individual leaders can work with their teams to improve psychological safety.

Artificial intelligence and psychological safety

Prof. Sanders: How does artificial intelligence relate to psychological safety? How will psychological safety evolve if AI takes over part of the work that people currently do?

Prof. Edmondson: AI is a two-edged sword. On one hand, AI can empower people in meaningful ways. It is a powerful tool that can help them with certain aspects of their job, to be more efficient, and may even build confidence as a result of that expanded capability. People can accomplish things they could not have accomplished before, or accomplish familiar tasks much faster, which can be empowering.

On the other hand, the broader uncertainty about AI – where it is going, which jobs and tasks it will replace, and which ones it will not, as well as what the timeline looks like – is creating generalised anxiety in workplaces. People worry about their roles, relevance, and future in ways that can undermine psychological safety if left unaddressed.

The only way to help address that generalised anxiety and strengthen psychological safety, rather than have it fly out the window, is by making it discussable. That means staying away from either overly negative or overly positive views and talking honestly about the uncertainty. The reality is that organisations really do not know where this is heading. This technology is changing faster than any technology organisations have seen before, which means it is creating more uncertainty than seen before. Leaders need to be humble about what is going on and encourage people to discuss it openly, express concerns, raise questions, and collaborate to figure out the implications together – rather than everyone worrying in isolation.

Learn more: Yoga for stress, therapy for bosses? Wellbeing programs miss the mark

Prof. Sanders: Do you find that discussions about AI and its implications for work become polarised, with some people very positive and others very negative, without enough evidence-based dialogue?

Prof. Edmondson: That is exactly what happens. People are either enthusiastic about or fairly negative towards AI. It becomes almost a binary discussion, akin to a birthday party conversation where people share their strong opinions rather than engaging with evidence. Some people say this will change the world entirely, and tomorrow, no one will have a job anymore. Others say it is no different than when organisations adopted laptops and the internet, and everything will be fine. Neither extreme is particularly helpful.

What is needed is much more data and much more evidence so organisations can have conversations that are grounded rather than opinion-based. The challenge is that the evidence is still emerging, and the technology is evolving so quickly that by the time studies are completed, the technology has changed. But organisations should still strive for evidence-based discussions rather than falling into evangelism or doomsaying. This polarisation is not helpful for organisations trying to navigate the actual implications of AI for their work, their people, and their strategies.

Prof. Edmondson recently presented at the World Business Forum in Sydney, where she discussed how to build psychological safety into everyday work and decision-making, along with key strategies for fostering learning, collaboration, and high performance. Main photo credit: Evgenia Eliseeva