Progress in the rear view mirror

The speed of AI and the lessons we are choosing to forget

There is a comforting narrative that appears whenever societies discuss major technological change. People say that we have seen this before.  They point to the Industrial Revolution, mechanisation, electrification, computers and the Internet.

They remind us that new technology always destroys some jobs but eventually creates others.
Then comes the reassuring conclusion.  “It all worked out in the end.”  The problem with this argument is that it quietly edits out the human experience of living through those transitions.  The Industrial Revolution may have transformed economies and increased prosperity over time, but for many of the people experiencing it, it was brutal.  Families were forced from rural communities into overcrowded industrial cities. Traditional livelihoods disappeared. Men, women and children worked in dangerous factories where health and safety protections barely existed. Children lost limbs in machinery. Workers died young. Entire communities lived in poverty while industrial wealth accumulated elsewhere.

We now look backwards and see progress.  At the time, many people saw fear, instability and exploitation.

That distinction matters enormously when we think about AI.  Because once again we are hearing the same reassuring language. We are told that jobs will change, new opportunities will emerge and society will adapt as it always has.

Perhaps it will.  But that does not mean the transition itself will be orderly, fair or humane.In fact, there are reasons to think this transition may be considerably harder.  Previous industrial changes unfolded over decades or generations. AI is developing at extraordinary speed. Entire categories of administrative, analytical and technical work are already being reshaped in months rather than years. Organisations are deploying AI tools faster than governments can regulate them and faster than education systems can adapt.

Yet most governments still appear to be treating AI disruption as a future issue rather than a current one.

That is understandable politically. No government wants to create panic. Most policymakers do not yet fully understand the technology. Economists themselves disagree about the scale of the impact.  So governments wait.

They wait for clearer evidence.
They wait for unemployment numbers.
They wait for social consequences that are impossible to ignore.

History suggests that by the time those signals become obvious, the disruption is already deeply embedded. This is not really a technology problem. It is a leadership problem. The real issue is not whether AI creates wealth. It almost certainly will. The real issue is who receives that wealth and who absorbs the disruption during the transition.

That question has existed in every industrial transformation.  Who gets the spoils and who gets spoilt?  During previous revolutions, the gains often concentrated rapidly while protections for ordinary workers emerged much later and usually only after political pressure, industrial unrest and social instability.

There is little evidence that this time will automatically be different.  If anything, the concentration of power may happen faster. A relatively small number of companies already control enormous computational capability, infrastructure and data resources. Meanwhile many organisations are approaching AI primarily through the lens of labour reduction and efficiency. That may improve short term profitability but it risks creating long term instability.

Because work is not simply about income. For most people, work provides identity, structure, meaning, status and social connection. If large numbers of people begin to feel economically unnecessary, the consequences will not just be financial. They will be psychological, political and cultural.

This is why the current conversation around AI often feels strangely shallow.  We spend huge amounts of time discussing capability and comparatively little discussing transition.

How do we retrain people fast enough?
How do we redesign organisations so human contribution still matters?
How do we prevent entire generations from feeling excluded from economic value creation?
How do we maintain social cohesion if productivity rises while employment security falls?

These are leadership questions as much as economic ones. And they need answering now, not after the damage becomes visible.  Because one of the great historical mistakes is assuming that technological progress automatically produces social progress. It does not.

Social progress requires leadership, preparation, intervention and sometimes restraint.  Without that, societies often discover too late that what looks like progress from a distance can feel very different when you are the person living through it.