10th August 2026
For decades, computers have been remarkably useful machines.
They have helped us write documents, manage accounts, communicate with customers and search enormous amounts of information.
But something fundamentally different is now beginning to happen.
The computer is starting to understand what we want done – and potentially do it for us.
That may prove far more important than the arrival of another chatbot.
From software tool to digital employee
Imagine sitting at your computer and saying:
"Research this story, check the latest figures, write an article in my usual style and put it into my publishing system."
Instead of opening several programmes and websites and carrying out each stage yourself, an AI agent could potentially coordinate the whole process.
The human provides the idea and makes the final decision.
The AI performs much of the work.
That is a profound change.
It means computers are moving from being tools that people operate towards becoming agents that carry out tasks on people's behalf.
And the same principle could eventually apply to millions of jobs.
The office could change dramatically
Consider an ordinary office.
Today there might be employees dealing with:
invoices;
spreadsheets;
appointments;
emails;
customer enquiries;
reports;
databases;
research;
administration;
document preparation.
Increasingly capable AI agents could potentially perform significant parts of all these jobs.
That does not necessarily mean millions of people will suddenly become unemployed.
More likely, many jobs will gradually contain less routine work and more supervision, judgement and human interaction.
But there is another possibility.
If one employee with AI can accomplish the work previously done by two or three people, a company may decide that it simply needs fewer employees.
And that brings us to the question that is rarely answered adequately.
Who gets the benefit?
Suppose a company employs 100 people.
New AI technology allows it to produce the same output with 70.
The company now has several choices.
It could:
A. Employ fewer people and increase profits.
B. Keep the workforce and produce more.
C. Reduce working hours while maintaining pay.
D. Increase wages because each employee is now considerably more productive.
E. Reduce prices and attempt to capture more market share.
Or it could use some combination of all five.
There is nothing automatic in economics that guarantees the benefits will flow to workers.
That is the uncomfortable part of the AI revolution.
Productivity could soar – but wages don't have to
This is where AI becomes much more than a technology story.
If productivity rises dramatically, the economy could potentially produce far more goods and services with fewer hours of human labour.
That could be wonderful.
Imagine a future in which people work 25 or 30 hours a week rather than 35 or 40, while enjoying the same standard of living.
But there is another possible future.
People work fewer hours because their jobs have disappeared, while the owners of the technology and businesses receive most of the gains.
The technology could create either greater prosperity or greater inequality.
The outcome will depend largely on how society chooses to distribute the gains.
This has happened before
The industrial revolution provides an important warning.
Machines dramatically increased productive capacity.
Over time, living standards rose enormously.
But the benefits were not distributed evenly at first.
There were periods of terrible working conditions, low wages and enormous inequality before political pressure, trade unions, education, welfare systems and changing labour laws altered the balance.
The computer revolution also transformed employment.
Millions of traditional administrative jobs disappeared or changed.
But millions of new occupations emerged around technology, finance, telecommunications, software and the wider service economy.
AI could repeat that process – but potentially at much greater speed.
The speed is what makes this different
Previous technological revolutions often took decades.
AI can spread globally almost immediately.
A new AI capability developed in California, London or Beijing can become available to millions of businesses within months.
That means workers may not have decades to retrain.
A person whose job changes dramatically in five years may need to acquire completely different skills while still paying a mortgage or supporting a family.
Governments therefore face a difficult challenge.
Education systems designed around preparing young people for particular occupations may need to become much more focused on adaptability and lifelong learning.
And then there is taxation
There is another enormous question.
Governments currently collect large amounts of tax from employment.
Income tax, National Insurance and consumption taxes all depend to some extent on people earning and spending money.
But imagine an economy in which companies can produce substantially more with substantially fewer employees.
Where does the tax revenue come from?
This could eventually reopen debates about:
taxation of company profits;
taxation of capital;
wealth taxation;
consumption taxes;
universal basic income;
negative income taxes;
shorter working weeks;
stronger social security systems.
None of these questions has an easy answer.
But governments may eventually have little choice but to confront them.
The winners could be enormous
It is important not to portray AI simply as a threat.
The potential benefits are extraordinary.
A small business could gain capabilities previously affordable only to large corporations.
A doctor could have AI assistance with medical information.
A teacher could receive help preparing personalised learning materials.
An engineer could have an intelligent assistant analysing designs.
A farmer could use AI to monitor crops, weather and machinery.
A journalist could research and prepare material much faster.
An individual could effectively have access to a personal digital assistant that would previously have required a team of employees.
AI could therefore democratise access to expertise.
That could create enormous economic opportunities.
But the transition could be painful
The problem is that economic benefits and economic disruption rarely arrive at exactly the same time.
A company can introduce AI tomorrow.
A worker displaced by that technology may need years to retrain.
A new industry may eventually create thousands of jobs, but those jobs may require completely different skills from the ones that disappeared.
That creates a dangerous gap.
And it is why the debate should not simply be:
"Will AI take our jobs?"
The better question is:
"How do we ensure that the productivity gains from AI improve ordinary people's lives?"
Perhaps the biggest change is still ahead
The arrival of AI agents running directly on computers could accelerate this process.
If AI remains something we visit through a website, its impact will be significant.
If AI becomes something built into our computers, phones, vehicles, workplaces and homes – quietly carrying out tasks on our behalf – the impact could be considerably greater.
The computer may become less like a machine we operate and more like an assistant that works continuously in the background.
And that raises a fascinating possibility.
Perhaps the biggest question about AI isn't whether machines will become intelligent enough to replace humans.
It is whether humans will become intelligent enough to decide what to do with the extraordinary productivity that those machines create.
Because if AI allows society to produce more with less human labour, we have a choice.
We can use the technology to make a small number of people enormously wealthy.
Or we can use it to create more leisure, higher living standards, better public services, shorter working weeks and greater economic security for everyone.
The technology is arriving faster than many expected.
The difficult part will be deciding who benefits from it.
Meta's Muse Glimmer Is AI About to Move From the Cloud Into Your Computer