AI and leadership: Do we remember to ask the right questions?

I think we have misunderstood part of the conversation about artificial intelligence. Far too often, we end up framing it as an either-or choice: either we use AI because it makes us more efficient, or we stay away from it because we cannot trust the technology, because we risk losing our jobs, or because it might make us less intelligent.

There are genuine concerns in that debate, and we need to take them seriously. But I find myself missing a different conversation: one about why we want to use AI, what we want to create with it, and how we can do so in a way that is responsible towards people, organisations and the society of which we are part.

This is not a question of technology. It is a question of leadership.

If we only look for savings, we may never find the opportunities

Imagine a CEO who discovers that AI can take over a range of tasks. That may well be the case. If employees’ work consists entirely of tasks that AI can replace, jobs may disappear. We should not dismiss that possibility.

But this is where a choice arises. Should the CEO focus solely on how many employees the organisation can do without? Or should the organisation also explore what might become possible if it invests in its employees’ skills at the same time as it invests in technology? I believe there is considerable potential in the latter.

Employees who are given new skills and time to experiment may discover new products, improve existing services or develop ways of working that management has not yet imagined. Not every organisation will be able to retain every employee, and not every experiment will succeed. But if we view AI solely as a replacement for what already exists, we risk using the technology of the future to preserve today’s idea of what an organisation should be.

Employees are not simply a cost that technology can reduce.
They can also be the people who discover what the technology makes possible.

Leaders need to set direction, not simply hand out access

The answer is not simply to buy an AI tool for everyone and encourage them to experiment freely. Quite the opposite. I believe that one of the first things that happened in many organisations was that employees were let loose with a new and fascinating technology without the organisation first establishing the necessary boundaries. That can create problems.

I would start by exploring the value AI could bring in relation to the organisation’s purpose and work. We cannot know everything in advance. Nor could we enter new markets or develop new products. We must form qualified ideas, try things out and learn as we go.

The organisation then needs to articulate its fundamental principles. Do we believe AI should be used alongside people and under human responsibility? Should its use contribute to better conditions for people and to sustainable development? How do we balance opportunities, costs, quality and responsible practice in each specific situation?

Then comes the practical framework: which systems do we use, and why? Where is data processed? How do we protect confidential information? Which tasks may AI support, who reviews its outputs, and where do we draw the line?

Only then does it really make sense to train employees and give them the freedom to work within those boundaries. Clear boundaries and freedom are not opposites. Boundaries can be precisely what makes freedom responsible.

We need to learn to think with AI, not just use it

One of my greatest concerns is not that AI necessarily makes people less intelligent. It is that it can make it tempting to skip the thinking altogether.

AI often acts like a mirror. The answers we receive are shaped by the questions we ask and the assumptions we bring with us. And we read those answers through the lens of our own state of mind. An anxious leader may ask a question in one way and focus on one answer. A curious leader may ask it differently and discover something entirely different.

That does not mean AI always does exactly what we ask. It can misunderstand, make mistakes and produce convincing answers that do not stand up to scrutiny. But our own questions, assumptions and emotions play a major role in the interaction.

That is why courses in technical skills and prompting are not enough. Leaders and employees need to develop critical thinking, critical dialogue and critical questioning skills. They need to be able to ask: what am I taking for granted? Which explanations have I overlooked? What evidence challenges my conclusion? And what do we actually know? For leaders, this becomes particularly important.

An AI thought partner that simply agrees with the leader can do more harm than good. It should be able to challenge, offer new perspectives and help the leader identify their blind spots. That requires a clear purpose, relevant information about the leader’s working patterns, a clear set of values and firm boundaries around what the system should and should not do. But responsibility for judgement and decision-making remains with the human being. A well-written AI analysis is not the same as an accurate understanding of an employee, a conflict or an organisation.

Leaders must have the courage to decide, and the courage to change their minds

Once the leadership team has made a considered decision to use AI, it cannot leave the organisation’s working methods entirely to each employee’s personal preference. Leadership means setting direction, even when we do not yet know the whole route. But that does not give leaders a monopoly on understanding reality.

If an employee can demonstrate that the chosen use of AI is reducing quality, or that another approach creates greater value for the task at hand, the leader needs to listen. A good leader has the courage to set a direction and the willingness to change it.

Much of the technical work can be delegated to specialists and consultants. What matters is that someone in the leadership team takes responsibility for the purpose, the values and the organisational consequences. To the leader who says that the demands of day-to-day operations leave no time to engage in AI, I am tempted to ask: can your organisation afford you not to get involved?

ft råd til, at du ikke involverer dig?

Who should benefit from the gains?

Let us say AI frees up working time. What do we do with it? Do we produce more? Lower our prices? Develop something new? Give employees greater freedom without reducing their pay?

The market may not need more production. And before deciding how to distribute the gains, we should ask whether the newly available capacity could be used to create new value. If not, the question of distribution becomes unavoidable.

Here, I would like to introduce a word into the AI debate that we do not use often enough: fairness.

If the organisation keeps all the gains, employees may reasonably ask why. If all the gains go to employees, customers may in turn ask why they are seeing none of the benefits. There is no single simple or correct formula for sharing the gains, but leaders should be able to properly justify their choices, taking into account the interests of the organisation, its employees and its customers.

I wish we were having a more fundamental conversation as a society about these principles. I worry that, in some places, short-term profit is being prioritised over consequences that we may only discover later. Until we have a broader shared understanding, organisations will have to take responsibility themselves.

My thoughts have found a voice

From my own experience, I know that AI can do something more than simply save time. I have always had many thoughts and ideas, but I have often held back from sharing them in writing. My spelling and grammar were a barrier. When I tried to express complex ideas or used figurative language, my message was not always understood in the way I had hoped.

With AI, my thoughts have found a clearer voice. I use the technology as an active editor. It can ask questions, challenge me, suggest structures and help me with language. It must not take over my ideas or change the fundamental assumptions behind the text. But it helps me communicate what I actually mean more clearly to others.

That is also why it concerns me when we begin to treat a text as less authentic simply because AI has helped to formulate it. You can still ask me what I mean. I can explain and defend the ideas. The technology has not replaced my voice. It has helped me use it.

And perhaps that is the best illustration I can offer of the opportunity I want leaders to recognise: not simply a technology that can do what we already do, only faster, but a technology that can help people realise potential that previously had no room to emerge.

We still need to remember the way back to other people. An insight developed through a long dialogue with AI does not automatically become a shared understanding among colleagues. We are responsible for making that translation too.

Let go of the fear, use your judgement and get started

I am not encouraging anyone to be naive. AI should be approached with healthy scepticism, clear boundaries and human accountability. We need to place demands on those who develop the technology and on ourselves when we choose to use it.

As leaders, we need to acknowledge our humanity: we can overlook connections, ask poor questions, be influenced by our emotions and get things wrong. That is not an argument against AI. It is an argument for learning how to work with it more intelligently.

Let go of the fear and replace it with sound judgement.
Acknowledge your humanity, approach AI with healthy scepticism and responsibility, and take the plunge.