By Ross Boyd, founder and director of Belfast-based, chartered accountancy, RBCA
We’ve spent more than a decade waiting for a productivity miracle. The UK’s productivity had been ailing since 2008, but new research from the ONS has revealed it’s grown by 1.1% since late 2024 – far above the official 0.2% estimate.
So, something is happening – and this is good news. But there is a temptation to join the dots quickly, and before we all prematurely praise artificial intelligence, we should examine things a little more closely.
AI adoption has certainly accelerated. So much so, the Organisation for Economic Co-operation and Development estimates that AI could eventually add between 0.4 and 1.3 percentage points to annual UK productivity growth. There is no doubt that AI will have a positive impact - but the word that matters here is eventually. The latest ONS evidence shows that roughly 35% of UK businesses use at least one AI technology. Yet, where businesses are using AI to improve operations, 63% report no change in company headcount.
READ MORE: From the Family Butchers to Blockchain: Kieran Kelly, Ubloquity
When the cost of labour increases, which it has, businesses want to find ways of using less of it. Employer National Insurance increased from 13.8% to 15% in April 2025. Improvements in efficiency are available across the board - a self-service till, accounting and booking software, an automated warehouse, better-designed internal processes. These suddenly become a worthwhile investment when employment becomes more expensive, but none of these changes require AI.
This makes it difficult to argue that AI has become the essential ingredient to a productivity boost. The composition of the workforce is an important area to examine.
Suppose a company eliminates ten relatively low-productivity jobs but maintains most of its output. Output per worker then rises. The business has become more productive, but that does not necessarily mean it has discovered a revolutionary technology, it may simply be driving efficiencies to increase productivity with fewer people. How can this be true? Well, productivity can rise because of effects on the labour market’s composition, including fewer lower-productivity workers and fewer entry-level jobs.
Productivity is measured across a broad set of sectors, rather than being confined to AI-intensive industries, and some organisations are openly shunning its value.
Barclays says there is no evidence to link productivity to its investment in AI, yet. Uber is reported to have reduced investment, after blowing its AI budget to little lasting effect. Klarna admitted recently that the advancements in AI had gone too far. It stopped hiring customer service roles, but started hiring again last year. Locally, hospitality businesses were out in force after the Fleadh commenting that we’ve reached peak AI slop with multitudes of posters, flyers, and food menus all looking woefully similar despite coming from different businesses.
Looking to the public sector, AI is certainly not a magic wand. In fact, it’s directly responsible for clogging up the system even further. There’s been a huge surge in complaints by consumers to schools, the PSNI, health trusts and councils, all fuelled by AI. It is a positive that AI is encouraging people to exercise their right to complain, but the other side of the coin is it is putting unmanageable strain on our already squeezed public services.
READ MORE: New AI and human-led QualiT app set to target the £21bn a year lost to avoidable construction errors
All of this matters because the weakest part of the current AI celebration is the assumption that correlation is causation. We see productivity rising. We see AI adoption rising. Therefore, AI must be responsible. But economics rarely present so perfectly. The current evidence more likely points to several forces operating simultaneously.
That’s the story we need to be telling. If we continue to position AI as a productivity miracle, while ignoring all the context, it will be a major mistake for growth. Some economists even believe that inflating its role could fuel a stock market bubble, which if it bursts, as many also predict it will, puts us at further risk of recession.
Higher productivity is the essence of growth, but productivity achieved by squeezing labour to its very limit could be a dangerous proposition. Let’s tread more carefully - AI is part of the solution, but not yet an answer to our productivity problem.
Read the Summer 2026 edition free online →
Stay connected with NI's tech community:
