Opinion: What if the AI boom really is a boom?

Finito World

For most of the past three years, the economic conversation about artificial intelligence has had a curiously apocalyptic quality.

AI was coming for the lawyers. It was coming for the programmers. It was coming for the copywriters, accountants, designers, journalists and, depending on which conference you happened to attend, perhaps everybody else too.

So it is worth noticing a rather different possibility emerging from the economic data: what if AI actually creates growth?

The latest UK figures offer a tantalising glimpse of that future. The economy expanded by 0.4 per cent in July, considerably better than economists expected. More interesting than the headline number was what lay underneath it. Services grew strongly and information and communication output rose 2.4 per cent in the month, with computer programming, consultancy and related activities expanding by 3.5 per cent. The Office for National Statistics has pointed to AI and related technologies as part of the explanation.

Nobody should hang out the bunting just yet.

Britain remains an economy with serious problems. Unemployment stands at 4.9 per cent, inflation remains uncomfortable, borrowing is expensive and households are hardly awash with confidence. One surprisingly good month’s GDP figure can disappear into a statistical revision almost as quickly as it arrived.

But there is something here worth celebrating nonetheless.

The great fear about AI has always been based on a fairly simple equation: if a machine can do something a human being currently does, fewer human beings will be required. At the level of an individual task that is plainly sometimes true. But economies don’t consist of a fixed number of tasks waiting to be divided between humans and machines.

Technology changes what it is possible to do.

A small company which could previously afford one designer may suddenly be able to produce the output of a creative department. A programmer can build in a week what once required a month. An entrepreneur can research a market, create a prototype, translate a website and analyse customer data without assembling a team of specialists first.

That does eliminate some work. But it also makes previously uneconomic activity economic.

And there are intriguing signs of precisely this phenomenon in America.

Indeed’s US data show software-development vacancies have risen by almost 15 per cent since the launch of Claude Code in February 2025, even while job postings overall fell by 7 per cent. The jobs being created are changing: senior and AI-fluent positions account for much of the recovery. Meanwhile AI is escaping the technology department altogether. The number of occupational categories in America with AI appearing in job titles has more than tripled since 2022, spreading into sales, HR, teaching, legal work, administration and customer service.

Even the physical infrastructure tells a more complicated story than the familiar robots-taking-jobs narrative. US data-centre job postings have more than doubled in two years. AI turns out to require not merely brilliant computer scientists but electricians, installers, engineers and construction workers.

Britain may be beginning the same transition. By June, 9.4 per cent of UK job advertisements mentioned AI or related technologies. Nearly half of data and analytics vacancies now do.

This matters because Britain desperately needs a productivity story.

For too long our national economic debate has consisted of governments arguing about how to divide a cake which stubbornly refuses to become much larger. Tax this. Subsidise that. Borrow here. Cut there. Each Budget becomes an increasingly elaborate exercise in moving money between pockets.

AI offers something altogether more interesting: the possibility of making the cake bigger.

That does not mean every consequence will be benign. Some occupations will shrink. Entry-level work may be particularly vulnerable. Workers will need to acquire new skills, and governments and employers will have to think seriously about how young people obtain the experience which yesterday’s graduates accumulated by performing precisely the routine tasks AI can increasingly perform today.

But we should be equally wary of making the opposite mistake: assuming every job performed differently must be a job destroyed.

The history of technological progress is largely the history of human beings finding more things to do once old things become easier to do. The spreadsheet did not abolish finance. The internet did not abolish commerce. The computer did not ultimately abolish programmers.

Perhaps AI will prove different. Nobody knows.

But July’s British GDP figures offer a small reason for optimism. For once, artificial intelligence has appeared in an economic story not primarily as a threat to employment, but as an apparent contributor to actual economic growth.

Four-tenths of one per cent is not a revolution. But revolutions in productivity have to show up somewhere first. Perhaps this is what the beginning of one looks like.