23rd September 2026
Much of the discussion about artificial intelligence and employment has become a modern tale of impending disaster.
Jobs will disappear. Young people will struggle to get started. Office workers will be replaced by computers. Perhaps, eventually, there will simply not be enough work to go around.
There are good reasons to take some of those concerns seriously. But there is another side to the story which can easily get lost.
AI is also a new general-purpose technology capable of making people considerably more productive. The question is therefore not simply how many jobs artificial intelligence might destroy. It is also what new work it might make possible, what existing jobs it might improve and what happens when goods and services can be produced more efficiently.
The International Labour Organization's latest research provides a useful corrective to some of the more dramatic predictions. It estimates that one in four workers worldwide are in occupations with some exposure to generative AI. Yet its conclusion is that transformation rather than replacement is the more likely outcome for most jobs. Only 3.3% of global employment falls into its highest exposure category.
That distinction matters.
A solicitor who uses AI to search thousands of documents is still a solicitor. An accountant using AI to examine a large dataset is still an accountant. A journalist using it to work through a lengthy report can spend less time on the mechanical work and more time checking what it actually means.
The same could happen in thousands of occupations.
A small business that could not previously afford specialist research, marketing or data analysis may suddenly have access to tools which can help it do those things. A sole trader could produce better quotations and invoices. A farmer could analyse information that previously required specialist help. A small Highland business might be able to compete with a much larger company in areas such as website content, customer enquiries and administration.
That is one of the less discussed possibilities.
Technology can allow a small organisation to behave like a much larger one.
There is also the productivity question. The OECD estimates that AI could add between 0.4 and 1.3 percentage points a year to labour-productivity growth in the more highly exposed G7 economies, including the United Kingdom, depending on how quickly the technology is adopted and how effectively it is used.
Productivity sounds like an abstract economic term, but it is really about how much can be produced from a given amount of work.
If a worker can do in four hours what previously took eight, that creates choices. The employer might reduce staff, but could instead produce more, reduce prices, expand into another market or allow workers to spend more time on tasks which machines cannot easily perform.
History suggests that technological change can do all of these things.
The printing press, mechanisation, electricity, telephones, computers and the internet all eliminated particular forms of work. They also created entirely new industries and made other activities economically viable.
That does not mean today's workers should simply be told not to worry.
The difference this time is the speed and breadth of the change. Earlier technologies often required factories, machinery, new infrastructure and years of investment. Generative AI can arrive as a software update. A person can start using it almost immediately.
That creates a genuine problem for workers whose existing tasks can be automated faster than they can acquire new skills.
The ILO's latest review says there is still limited evidence of large-scale job displacement, but it also identifies risks including weaker opportunities for younger workers and growing inequality.
There is another danger. The benefits may not be shared automatically.
A large company with money to invest in AI, good data and skilled employees may gain much more than a small business struggling with costs and skills shortages. The OECD also points to barriers for smaller firms, including the cost of adoption, infrastructure and lack of suitably skilled workers.
So the sensible position is neither "AI will destroy employment" nor "AI will make everybody better off".
It is a technology capable of doing both.
Some jobs will disappear. Some will be transformed. New jobs will emerge that are difficult to imagine today. And millions of existing workers may simply find that AI becomes another tool they use every day.
The biggest economic question may therefore not be whether AI creates or destroys jobs.
It may be whether society can adapt quickly enough to make productive use of the technology while helping the people whose particular skills are becoming less valuable.
That is a much less dramatic story than the robot taking everybody's job.
It is also probably closer to the way technological change actually works.