Brickchat

Sitting in his office, I would often find myself wondering why he had asked me in there. As we entered that morning’s fifteen minute of discussion, we had covered everything from family dynamics, to home maintenance, and gardening. Anything other than what I went in there for.

The site outside was safely up and running for another day’s operation. Last minute deliveries, labour issues and hook time quarrels resolved, his focus had drifted to life outside the hoardings. I enjoyed those mid-morning lulls, but one time he became reflective. Fixing me with a stare he proclaimed that,

“The only way that any of us get paid in this business is by putting a brick on top of a brick.”

Most knew him as the marauding figure of strength that he projected. Overbearing, authoritative. He was derided and misunderstood. Feared by some, tolerated by others. Towards the end of our time working together I would often, sometimes publicly, rebuff his instructions, and criticise his methods. Sometimes I knew better and other times not. With or without my agreement he always found ways to keep a job moving forwards in difficult circumstances.

We had always viewed the work differently and it was better as a result. His focus on deadlines drove him and the projects forward at such a pace that I had to choose. Either work with him or be left alone in my site office chasing after what I thought we should have done.

This tension is a constant in construction. It is also one of the biggest arguments in favour of multi-generational teams. In knowing what it took, brick by brick, to start, progress, and finish a building, he knew the dangers of prevarication. Paired with my own command of technical and legislative requirements, we were effective. Knowledge is nothing without action. I needed him just as much as he needed me.

It is in remembering those mornings, watching the machinations of the site and those on it, that I bristle when hearing about the role of technology and the supposed changes afoot.

He isn’t paying attention, he does care, because he has to open up more work areas for the concrete guys or they are walking off site. He needs to deal with a bad batch of bricks that we have been charged for, as well as the thermal store tanks that have been delivered a month too early. And let’s not forget the dressing down of an errant scaffolding gang caught misbehaving in sight of the public.

Most of the crystal-balling around AI is coming from the software companies buying up hardware and engineering resource to serve whatever need might exist in the future, evidenced by Autodesk sponsoring a round table for RIBAJ (June 2024). It followed the predictable format of drafting in tech savvy designers to talk in broad strokes about the potential of AI, with no real outcomes other than column inches filled.

In watching this I ask, did the last round of digitisation make construction a safer, fairer place? The Grenfell Tower project was planned, procured and built in the midst of this post-analog age. Mistakes, omissions, and disputes occurred in spite of the existence of pdfs, dwg and doc file formats. Projects the world over continue to be badly run no matter what manner of tools people have at their fingertips.

So will AI change everything ? NO. It’s a tool to be used. Can it be touted as anything more than this in a world that is inherently physical? AI might be able to tell us where and when to replace a valve in a hot water system but you still need a plumber to go to the property, isolate the area, cut the old fitting, replace it, pressure test and re-commission.

The ability to filter large volumes of information and exercise judgement is another big ticket item touted as an area for innovation. In spite of this speculation the responsibility for taking action in complicated circumstances is where the human will remain essential. How does one create meaningful training data for an AI working within a construction site? Electric vehicles have millions more data points and its collection extends to anyone who has forgotten a password while online shopping. Not so much in construction.

Is AI at risk of becoming a blockchain-type fallacy where we are told that the sharing of information and knowledge in an open ledger creates great excitement only to slam into the harsh reality of survival in one of the highest-risk lowest-margin business models on the planet. There is no Creative Commons when it comes to collecting training data from construction sites. Would Laing O’Rourke be happy to share data in a pool used by their closest competitors? If not we are reliant on parties developing their own systems and data sets. No different to the faltering we have seen in the world of Whole Life Carbon assessments.

Findings from a recent Goldman Sachs research report are telling. Management, Technical Advisory and Professional roles occupy the largest groups likely to be affected by the rise of AI. Construction and Extraction are among the least affected.

Super Users, the book by Randy Deutsch contains much speculation about AI’s applications in design iteration. Reducing its potential to this alone risks designers being relegated to the role of machine operator. The precedents and design studies to date oscillate back to Architecture as the standalone sculptural form.

No mention is made of the city grid, its infrastructure or zoning. Nor how changing population demographics or economic patterns could be used to masterplan our cities. Instead we are obsessing over case-by-case unitised facade iterations, a practice that in a century’s time will be (I think) deemed an irrelevance at best, and climatically nonsensical at worst.

Material science. Product development. AI can be and is being used to simulate compounds for increased electrical conductivity, fire resistance, and more. Surely more worthy uses for a technology reliant on good training data.

A building should be judged by the quality of its installation, its commissioning and long term maintenance. Regulatory work is ongoing to encourage this focus from industry. Until legislation and normative responses to same are fully developed judgment will be required in the absence of data.

The chasm that exists between the building that is designed and the one that is built will remain unless one discipline can deliver a project end to end in-house. Until then, the people we rely on to deliver our construction projects must keep moving forwards, brick by brick.