Living inside the machine

I first heard Psycho Killer on Radio Luxembourg.

I was growing up in Yugoslavia, and at night I would listen to the station on my parents’ old Loewe valve radio. The signal travelled a long way to reach us. It faded in and out, occasionally disappearing beneath static before returning again. You had to listen carefully.

Radio Luxembourg offered a glimpse of another world. Music arrived from somewhere beyond the borders of the one I knew, carrying with it different voices, different ideas and different possibilities. I doubt I understood much of what David Byrne was singing about when Psycho Killer came through that radio. What I remember was the sound. The nervous voice against that relentless bass line. Something human, awkward and unsettled held within an extraordinarily controlled rhythm.

At the time, it was simply a song I loved.

Only much later did I begin to recognise something else in it.

The comfort of the rhythm

There is something strangely comfortable about Psycho Killer.

The observation sounds wrong at first. The voice is anxious. The language fragments. Something about the person speaking seems unable to settle comfortably within the world around him. Yet underneath all of this, the rhythm remains remarkably certain.

The bass repeats.

The structure holds.

Whatever is happening to the human being inside the song, the machinery keeps moving.

Perhaps we can live with considerable uncertainty when something else provides the rhythm.

I have found myself returning to that thought as artificial intelligence has begun to occupy more of our organisational and everyday lives. Much of the discussion around AI concerns capability. We ask what the technology can do, what it will soon be able to do and which parts of human work it might eventually replace.

Increasingly, I find myself interested in a different question.

What happens when living inside the machine becomes more comfortable than living outside it?

Making the world manageable

Organisations have always needed to simplify the world.

Customers become segments. Experience becomes data. Relationships appear as organisational charts. Culture is expressed through values. Behaviour becomes metrics. Uncertainty is translated into risk.

These abstractions are useful. Without them, much of organisational life would become impossible to navigate. They allow us to find patterns in complexity and to share an understanding of things too large for any one person to experience directly.

But every abstraction leaves something behind.

A customer segment cannot tell us what it felt like to stand at a railway station after receiving difficult news. An engagement score cannot capture the hesitation before somebody decides whether it is safe to speak. A process map cannot reveal the years of trust between two colleagues that allow a broken system to continue functioning. A performance metric may tell us what happened while leaving invisible the circumstances that made the behaviour entirely reasonable to the person involved.

The abstraction works because it contains less of the world.

And less of the world is often easier to manage.

Data is easier than humanity.

The machine becomes convincing

Artificial intelligence extends our capacity for abstraction enormously.

It can synthesise thousands of documents, identify patterns across datasets, summarise conversations, classify behaviour and generate plausible responses almost instantly. Increasingly, it can give us something that resembles understanding without requiring us to experience the slow, uncertain and sometimes uncomfortable process through which human understanding develops.

This is an extraordinary capability.

Its usefulness is precisely what makes it interesting.

A summary can begin to feel like having read.

A sentiment score can feel like having listened.

A customer model can feel like having understood.

A recommendation can begin to feel like judgement.

None of these substitutions needs to happen deliberately. They can emerge gradually through convenience. The representation becomes sufficiently useful that the distance between it and the thing it represents becomes harder to notice.

We begin to inhabit the abstraction.

Outside the machine

The world beyond it remains inconvenient.

People contradict themselves. Conversations wander. Stories refuse to fit neatly into categories. Someone says something that makes little sense until we discover what happened three years earlier. A behaviour that appears irrational on a dashboard becomes entirely reasonable when we stand beside the person doing the work.

Sometimes we spend an afternoon observing, listening and asking questions and return without a conclusion.

There is friction in experiences like these.

Yet friction is sometimes where understanding begins.

Attention asks us to remain with the world for slightly longer than might feel efficient. To notice before categorising. To listen before summarising. To allow an observation to disturb what we thought we knew.

Sometimes nothing happens immediately.

Then, much later, something connects.

Judgement develops this way.

It grows through experience, context and memory. Through conversations whose significance becomes apparent only afterwards. Through encountering perspectives that unsettle our own. Through noticing the distance between what a system says should happen and what people actually do.

The process is difficult to optimise because we rarely know in advance which observation will matter.

A more human question

This is why I have become less interested in whether artificial intelligence will eventually reproduce human intelligence.

There is another question closer to hand.

What happens to human judgement when we stop practising the conditions from which it emerges?

As machines become increasingly capable of processing information, recognising patterns and producing answers, our contribution may depend less on competing with those capabilities and more on deciding where our attention should go.

• What matters here?

• Whose experience is absent?

• What has been simplified?

• What are we no longer seeing?

• What should be preserved?

• What deserves to be done?

These are not questions that require us to reject the machine.

They become more important because the machine is so useful.

The seductive interior

There is another reason living inside it may become attractive.

Increasingly, technology surrounds us with versions of the world organised around what we already know. Recommendations anticipate our preferences. Search anticipates our questions. Algorithms decide what might deserve our attention. Generative systems can give form to an idea before we have fully discovered what we think.

The experience can be remarkably frictionless.

Outside that environment, the world remains stubbornly uncurated.

We encounter people we did not choose. Ideas that do not fit our assumptions. Places that make no attempt to anticipate our preferences. Conversations without an obvious purpose.

The world interrupts us.

And interruption is one of the ways our field of view widens.

Perhaps the challenge is not to choose between these worlds, but to retain our ability to move between them. To benefit from abstraction while remembering what has been abstracted. To enter the machine when it helps us see, and to know when to step outside it again.

Listening for the signal

Perhaps paying attention in the age of AI is partly about remaining receptive to those signals. Not because the machine is somehow opposed to being human, but because no representation of the world, however convincing, can contain all of the world.

The danger may not be that artificial intelligence eventually becomes indistinguishable from human intelligence.

It may be that life inside the machine becomes so comfortable that we stop noticing how much of being human happens outside it.

This question sits at the heart of my new book, Attention: A Field Guide.

I wrote Attention because I have become increasingly convinced that judgement will matter more, not less, as machines become more capable. AI can help us process information, recognise patterns, synthesise evidence and generate possibilities at extraordinary speed. But judgement asks something different of us. It asks what matters, whose experience is absent, what deserves to be preserved and, ultimately, what deserves to be done.

Those questions begin with attention.

Attention is a field guide for practice. It brings together a collection of practices developed through years of working inside organisations, designed to help us widen our field of view, notice relationships that might otherwise remain hidden and develop the judgement needed to navigate uncertainty with greater clarity and care.

It is not a methodology to follow from beginning to end. It is designed to be carried, returned to and lived with. A trusted companion for those moments when the answer seems obvious, the data appears convincing, or the machinery of an organisation is moving so efficiently that we might forget to look outside it.

Perhaps that is why I keep returning to that old radio.

Somewhere in my memory, the bass line is still repeating. Somewhere behind it, through the static, a voice is trying to get through.

We just have to keep listening.

Attention: A Field Guide is available now on Amazon.