The AI questions have changed
For two years, almost every conversation I had about AI started in roughly the same place: where are the opportunities? What are the priority use cases for us? How fast can we take them? Who is ahead?
That conversation has not gone away, but it is no longer the only one.
A friend of mine, John Metselaar, who leads The Conference Board’s Global Innovate & Transform Institute and spends much of his time with senior leaders across industries, told me that in recent months, he has noticed two questions appearing more often in those conversations. Neither is really about opportunity.
The first question is personal.
Leaders are starting to experience AI not just as something they deploy, but as something they personally have to lead. Can I do what is being asked of me? How am I supposed to lead this when I don’t know where it is going? And what am I actually leading towards?
A consultant can map use cases. A vendor will happily create you a rollout roadmap. But whether you can keep a team together through a change whose destination is still unclear is your problem. And, as a leader, you are still expected to appear reasonably confident while figuring it out.
A survey published this spring found that 73 percent of CEOs reported stress or anxiety about their company’s AI strategy.1 It is a vendor-sponsored survey, so I would not put too much weight on the number, but the direction feels correct. It is also what John is seeing in his conversations.
The second question is even harder.
Interestingly, John and I are both seeing more and more leaders ask themselves questions that have not been part of management conversations before:
What will people do when a growing share of the work is done by machines? What will we be valued for as employees and humans? What will make someone feel that they have done something worthwhile at the end of a day? At the end of a life?
Not long ago, those questions belonged mostly to philosophers. At the current pace of AI development, they are fast becoming management questions, with very concrete consequences for the business.
Researchers have described how AI is taking over many of the small, checklist-style tasks that used to give people a sense of progress: emails answered, documents processed and summarised, analyses completed.2 People may end up doing more while feeling that they have accomplished less. Some may even feel a kind of guilt about no longer doing parts of their jobs themselves.
I don’t think that is simply a productivity problem. There is something deeper going on. With the typical AI productivity hacks, we are removing some of the everyday achievements that used to tell people their work matters. That is probably already affecting their motivation.
That is part of the reason why I am not optimistic about the next few years, without even going down the rabbit hole of doomsday scenarios. It seems inevitable that this will get worse as AI moves from automating small tasks here and there to replacing ever larger parts of people’s jobs.
Another perspective
I’m grateful to John for pointing me towards a counterweight: a book he recommended, Angus Fletcher’s Primal Intelligence. Although I wouldn’t recommend that book without reservation, I found it uplifting. It argues that four deeply human capabilities (common sense, intuition, imagination and emotion) cannot be replicated by machines, because they are not computations that can be calculated and statistically calibrated, but physiological processes rooted in our biology. Common sense, for example, involves knowing when you do not know, which, Fletcher argues, an AI cannot recognise by itself. Some of the neuroscience in the book is shaky, to say the least, and the constant anecdotes from the US military really annoyed me. But that particular idea has stayed with me, and has given me something to think about and hope for.
In an AI transformation, admitting that you don’t know is not a failure. It may be one of the few things a leader can model for their people that the technology cannot.
What then?
The two questions we’ve raised are different in kind: the first is about the leader’s ability to navigate the change and the second is about what the change means for the people they lead, and for themselves. But they meet in the same place: the person at the front of the room, whom everyone in the team looks to for guidance, and who is at the same time part of the situation.
People watch what a leader does when there is no obvious answer. They notice whether uncertainty produces panic, denial or thoughtful action. And most leaders, privately, are dealing with the very same uncertainty and doubts as their employees.
There is some evidence that leadership changes how people experience AI. Gallup data from the US suggests that workers who use AI more frequently are more concerned about losing their jobs, except where they experience well-structured management.3 The technology may be the same, but the experience of it isn’t.
In practice, none of this is particularly glamorous. You name the question instead of rushing to close it, and admit to your team that you do not know the answer, at least not yet. You tell your team what you know now, and what you don’t. You are clear about how you will make decisions under the uncertainty, even when you cannot yet say exactly what those decisions will be.
That is difficult when the board wants a roadmap, regular execution and status updates, and the employees want certainty. But dealing with uncertainty is a profound leadership capability, and like most capabilities it can be practised only in the hardship of the situation itself. Meaning: now.
I would even argue that leading under uncertainty is probably the most essential leadership skill of them all.
This is where I think leadership work becomes interesting: it is not a leader’s job to produce yet another AI strategy. Their value lies in helping the people who have to carry the uncertainty in daily operations to see a perspective that goes beyond the fear of losing their jobs. It will also help them stay useful and find purpose, even if their jobs are genuinely under threat.
The opportunities and use cases belong in strategy decks. They are the easy part.
Unfortunately, the hard questions are for the people who must lead the change.
