The next leadership challenge isn't how much intelligence we can deploy.
It's how much human attention we can meaningfully sustain.

I've noticed something about working with AI.

It doesn't get tired.

It doesn't need a coffee, lose concentration or suggest we pick things up tomorrow.

It just keeps going.

Another question. Another possibility. Another connection. Another piece of work.

And because it's often genuinely interesting, so do I.

I've had occasions when I've been working with AI for hours, exploring ideas, making connections and creating things that would previously have taken days.

Then I've noticed something.

My head feels numb.

I'm still working. Still asking questions. Still getting answers.

But I'm not really thinking very well anymore.

The AI is as productive as ever.

I am not.

And I suspect this experience tells us something important about the future of leadership.

What happens when we remove the friction from work?

For most of our working lives, friction has been something we've tried to eliminate.

Waiting for information. Organising meetings. Finding people with expertise. Preparing reports. Getting decisions made.

AI is removing enormous amounts of that friction.

And much of that is wonderful.

We can explore possibilities faster, access expertise more readily, challenge our assumptions and create things that previously seemed beyond our reach.

But I've begun to wonder whether some of the friction in our working lives was doing something useful.

A meeting ended.

A colleague needed time to think.

A report took a few days to prepare.

We went home.

Not all of those interruptions were wasted time.

Sometimes they created the space in which we reflected, reconsidered, made connections or simply recovered the capacity to think clearly.

What happens to human contribution when the work no longer has natural stopping points?

And what happens when that isn't just one person working with AI, but an entire organisation working with teams of AI agents?

We may be solving one bottleneck and creating another

Imagine a leader with a team of 12 people and 20 AI agents.

The agents can research, analyse, monitor, prepare recommendations, generate reports and carry out tasks continuously.

Work that once took days can be completed in minutes.

Fantastic.

Except that much of this work creates something else.

Information to absorb.

Recommendations to evaluate.

Exceptions to investigate.

Decisions to make.

Risks to consider.

And people who need to understand what is happening.

The organisation has dramatically increased its capacity to produce work.

But has it increased its capacity to understand, judge and act upon that work?

Not necessarily.

We may have moved the bottleneck from production to human attention.

And unlike AI processing capacity, human attention cannot simply be scaled by adding more computing power.

This is where I think leadership needs to change.

Intelligence Allocation: which intelligence should do what?

I've been developing a concept within my HI:AI® work that I call Intelligence Allocation.

It is the leadership discipline of deliberately orchestrating human and artificial intelligence.

Not simply deciding which tasks to automate.

But deciding which intelligence should undertake which work, how they collaborate, and where judgement, decision authority and accountability should sit.

The question isn't just what AI can do.

It's what AI should do.

And what humans should continue to do, experience and learn.

Because there's another danger in automating work.

We may remove the very experiences through which people develop expertise.

If AI undertakes the analysis, prepares the recommendation and suggests the decision, what opportunities remain for a less experienced person to develop their own judgement?

We might improve the immediate output while quietly diminishing the human capability needed for tomorrow.

That's not progress in my book.

It's why Intelligence Allocation must consider both the work being done and the human being developed.

But I've realised that Intelligence Allocation on its own isn't enough.

Attention Allocation: where should human thinking be invested?

The second concept I've been developing is Attention Allocation.

This is about something different.

Where should finite human attention be invested, and how do we protect the capacity to exercise it effectively?

I see two equally important responsibilities.

The first is about value.

If AI can undertake more of the routine analysis, administration and production, where does human attention make the greatest difference?

Perhaps in asking the question nobody else has asked.

Noticing something that doesn't quite fit.

Understanding how a decision will affect people.

Connecting ideas from completely different fields.

Exercising judgement when there isn't a clear answer.

Or imagining something that doesn't yet exist.

That's where I want human intelligence to be invested.

But the second responsibility is about sustainability.

How much attention can a human being meaningfully sustain?

Because there's little point in directing people towards higher-value thinking if we simultaneously overwhelm them with more information, more decisions and more demands for their attention.

An exhausted person doesn't necessarily make better decisions simply because AI has done the preliminary work.

And a person who is still responding isn't necessarily a person who is still thinking effectively.

That's the distinction I've experienced myself.

AI can keep producing long after the human has stopped thinking well.

The illusion of productivity

I think this deserves particular attention.

When we're working with AI, it's possible to feel extraordinarily productive.

There's always something happening.

Something new to read.

Another angle to explore.

Another answer arriving.

And sometimes that really is productive.

But sometimes we're simply continuing.

We're stimulated by the interaction without necessarily creating much additional value.

I've started to recognise that point in myself.

When another answer is no longer helping me think better.

When I'm accumulating possibilities rather than making progress.

When the most useful thing I could do is stop.

And yet the temptation is to keep going because AI makes it so easy.

I wonder how many leaders are beginning to experience something similar.

And how many will unintentionally design that experience into the working lives of their people.

Removing friction from production must not mean removing the space humans need to think.

The Conductor: a different role for leadership

I've previously described the emerging role of an AI-era leader as The Conductor.

I like the metaphor because an orchestra doesn't create something extraordinary simply by having more instruments.

Nor does every instrument play continuously.

The conductor understands what each can contribute.

When it should contribute.

What needs to be heard.

What needs to be quieter.

And how the different parts come together to create something none could achieve alone.

I think leadership in an AI-enabled organisation is becoming something similar.

The leader must conduct human intelligence, artificial intelligence, human attention and decision authority.

That means making deliberate choices about both Intelligence Allocation and Attention Allocation.

Which work belongs with people?

Which belongs with AI agents?

What can happen autonomously?

What genuinely requires human judgement?

Where should people invest their attention?

And how do we prevent the sheer volume of AI-generated work from overwhelming the people responsible for it?

The aim isn't to have every instrument playing all the time.

It's to create something valuable from the intelligence available.

And to ensure the humans involved can sustain their contribution.

A human in the loop isn't necessarily a human in control

There's another reason I think this matters.

We're rightly concerned about maintaining human oversight as AI becomes more capable.

But what does that actually mean?

Imagine an AI agent produces 50 recommendations requiring approval.

The responsible manager has 20 minutes before the next meeting.

They approve 48.

Technically, there was a human in the loop.

But how much meaningful judgement was exercised?

Did they understand the recommendations?

Could they identify what was missing?

Did they have sufficient time to challenge the conclusions?

Were they genuinely in control?

Or were they simply the final step in a process moving faster than their capacity to evaluate it?

I've said before:

A human in the loop is not protection if the human no longer understands, questions or meaningfully controls the loop.

This connects to another concept I've been developing, Agency Gap®.

The possibility that people become increasingly proficient at using AI while losing opportunities to exercise independent thought, judgement and meaningful control.

Attention overload could be one mechanism through which that happens.

We need more evidence to establish the relationship.

But I think leaders would be wise to investigate it now.

The next leadership challenge

For years, we've measured productivity through output.

How much work can be completed?

How quickly?

At what cost?

AI is changing the answers dramatically.

But perhaps we need to change some of the questions.

Are our people thinking better?

Are they developing greater capability?

Do they have sufficient attention available for the decisions that matter?

Are they still able to question, challenge and exercise independent judgement?

Are they becoming more effective — or simply more occupied?

And can they sustain that contribution without sacrificing their well-being?

These aren't soft questions.

They're questions about the quality of leadership, organisational performance and the future of human capability.

Within Human Frontier Thinking®, my starting point has always been that there's more inside people than they know.

AI offers extraordinary opportunities to discover and develop that potential.

But only if we design the relationship between humans and AI deliberately.

We should be using artificial intelligence to expand human capability.

Not exhausting human beings in pursuit of greater machine productivity.

The leadership challenge is no longer simply how much intelligence can be deployed.

It is how much human attention can be meaningfully sustained — and where that attention creates the greatest value.

And perhaps the question every leader should now be asking is this:

If AI is removing the friction from work, what are we deliberately putting in place to protect the human capacity to think?

---

Carol Glover | Founder, Human Frontier Thinking® and HI:AI®

Exploring the future of human capability, leadership and human–AI collaboration.

#HumanFrontierThinking #HIAI #FutureOfLeadership #HumanCapability #AILeadership