Undoubtedly AI is shaping the way we work but, as with we need to keep things in perspective – a perspective that should be data driven, writes Richard Steer
Every decade or so, it feels as though someone announces that a new technology is coming to sweep away half the jobs in the built environment. The latest contender for this role is artificial intelligence (AI). It is a huge change-maker, true, but are the claims of it delivering job-loss Armageddon, at least in the short term, exaggerated?
UK unemployment hit a near five-year high last month and, since the launch of ChatGPT in November 2022, unemployment has ticked up across much of the industrialised world. A chorus of commentators has been quick to blame AI and Kristalina Georgieva, head of the IMF, even warned of a “tsunami” of technology‑driven job destruction.
It is a compelling narrative. It is also – at least for the built environment – far too simplistic.
The US workforce is six times larger than it was in 1900, despite – or perhaps because of – a century of mechanisation, automation and digitisation
AI is important; it may even herald a new industrial revolution, but we need to keep things in perspective – a perspective that should be data driven.
History tells us that technology always destroys some jobs. It also tells us that it creates others, often in greater number. The US workforce is six times larger than it was in 1900, despite – or perhaps because of – a century of mechanisation, automation and digitisation.
Employment has risen through the arrival of the personal computer, the internet, robotics and BIM. The idea that AI alone will break this pattern and usher in a “new era of mass unemployment”, as London mayor Sadiq Khan put it, deserves a more sober assessment.
The fear is that AI is different – that it can replace not only routine tasks but also the creative, judgment‑based, interpersonal work that has traditionally been the preserve of managers, consultants and professionals. If that were true, the built environment would be in the firing line.
Project managers, planners, cost consultants, architects, engineers – all roles that involve synthesising information, making decisions and communicating with stakeholders. If AI can do all that, what happens to the people?
But before we start rewriting job descriptions, it is worth looking at the data. Yes, unemployment has risen since late 2022 in the UK, US, Germany, France and elsewhere. Yes, several studies do show a weakening in employment for younger workers in AI‑exposed occupations, such as software engineering. A Stanford analysis of private‑sector payrolls even found “early, large‑scale evidence” that generative AI is disproportionately affecting entry‑level workers.
However, much of the labour‑market weakness can be explained by something far more old‑fashioned, and that is cost control. Inflation surged in 2021-22. Central banks responded with the sharpest interest‑rate rises in decades.
The US price level is 24% higher than in early 2021, with interest rates jumping from 0.25% to 5.5% in little more than a year. That alone would have been enough to make most employers cautious.
John Burn‑Murdoch of the Financial Times analysed millions of job ads and found that the hiring slowdown began six months before ChatGPT appeared – precisely when interest rates were rising and equity markets were falling. Tech and professional services – the sectors most exposed to AI – are also the most sensitive to interest‑rate shocks. And in tech, many of the job losses simply reversed the pandemic hiring binge.
Youth unemployment is another red herring. Young workers are always the first to feel any cooling in the labour market. Employers cut hiring before they cut staff, so the pain shows up in the under‑25s long before it reaches the rest of the workforce.
None of this means that AI is irrelevant, of course. It simply means that its impact so far is more modest when compared with the constant churn caused by older technologies.
Self‑service checkouts have reduced the number of retail assistants by nearly 10% since 2023. Online gambling wiped out 62% of betting‑shop managers between 2022 and 2024. Online banking cut bank‑clerical roles by 28% between 2015 and 2017. Technology reshapes the labour market all the time. AI is just the latest addition to the mix.
So, what does this mean for the built environment?
First, management roles will change but not disappear. AI can already produce draft reports, risk registers, programme analyses and cost plan comparisons. It can summarise meeting notes, flag contractual inconsistencies and generate early‑stage design options. But it cannot walk a site, sense when a subcontractor is struggling, negotiate a compromise between two warring stakeholders or persuade a client that the impossible is, in fact, impossible.
The built environment is a messy, human, political space. AI is a tool, not a substitute for judgment.
Second, trade roles may actually be boosted. The more digital the industry becomes, the more valuable hands‑on skills become. You cannot automate bricklaying on a Georgian terrace refurbishment. You cannot outsource M&E installation to a chatbot.
AI is not going to wire a distribution board or fix a leaking soil stack any time soon
If AI accelerates design and management processes, it may increase the volume of work reaching site, and therefore the demand for skilled trades. The industry already faces chronic shortages of electricians, plumbers, carpenters and groundworkers. AI is not going to wire a distribution board or fix a leaking soil stack any time soon.
Third, the biggest risk is arguably not redundancy but complacency. AI will reward those who learn to use it and punish those who ignore it.
The professionals who thrive will be those who treat AI as an amplifier – a way to work faster, test more options, spot more risks and communicate more clearly. The tradespeople who thrive will be those who combine craft with digital literacy – reading 3D models, using augmented‑reality overlays, interpreting sensor data and working alongside increasingly automated equipment.
A recent study by the Yale Budget Lab found that the rate of job‑mix change in the US since the launch of ChatGPT is no different from the rate seen after the arrival of the PC or the internet. In other words: not a revolution, but another turn of the wheel.
AI will reshape the built environment. But it will do so gradually, unevenly and – crucially – with humans still very much in the loop. The robots are not coming for your hard hat just yet.
Richard Steer is chair of Gleeds Worldwide and a Building The Future Think Tank commissioner
No comments yet