We are standing on the brink of an AI revolution that will reshape not only professions, organisations and projects but also our day-to-day lives. James Bowles of Freeform explores the potential impact of large language models across the construction sector
Imagine sometime in the near future. You return to your desk after a holiday and ask, “Morning Jarvis, any updates? Where should we focus this week? Who should I be catching up with in priority order?”
It’s an understatement to say that AI is about to change our professional lives. Recent leaps forward in AI have captured our collective imaginations.
With the release of tools like ChatGPT and Midjourney, it’s clear that some AI tools are migrating from research labs to our daily lives. The exponential adoption, exciting new features (seemingly pushed daily), and a glimpse at the huge opportunity have added to the hype. We now stand on the edge of what is arguably a deeper technological revolution than the printing press, computers or the internet.
And what will this mean for our industry? How will “the great intelligence unlock” change our professions, institutions, organisations, and projects?
The future is coming at us, fast
With every emerging technology or innovation comes the inevitable question, “it looks awesome, but what problem is it solving?”
AI is simply too big to pin to a particular problem, or set of problems, in our industry. It’s the same as asking “what problem will the internet solve for our industry?” back in 1980.
If we force ourselves to be reductive, we could say the internet speeded things up. And the same will be true of the AI revolution that is about to sweep over us. It will speed pretty much everything up.
Throw in exciting developments in quantum computing, and we’ll soon have exponentially more power and compute speed. It will speed up communication, improve collaboration and automate large parts of projects and organisations.
But what does this mean on the ground and what should we be preparing for?
It’s hard to predict exact timelines, but change is coming soon. Think months and years, not decades. One thing is certain- the magnitude of change will be significant.
To keep things focused, I’ll concentrate on one type of AI: large language models (LLMs), such as Google’s Bard, Meta’s Llama, ChatGPT or Anthropic’s Claude.
What industry issues could AI improve or resolve?
- We have a widening skills gap.
- Our projects are getting more complex and challenging.
- Productivity has remained stagnant for decades.
- Sharing knowledge on projects and within organisations is often slow, inconsistent and manual. This applies to both internal project knowledge (related to design, contract etc) and external references (industry regulations, standards etc).
- Miscommunication and misalignment are common, whether among individuals, across disciplines or between businesses. This is especially true on large, complex projects.
- The growing, dynamic corpus of project language and information makes it harder to read the signal from the noise. Decision-making suffers.
The structured use of LLMs to address this set of problems will be transformative. Imagine centralised models that can combine all project knowledge and information, including contracts, drawings, reports, meetings, cost plans, programmes and emails.
These “project co-pilots” would grow mature with the project and be extremely useful as a support to the project team, and good for onboarding new members.
This single resource could then be used regularly by anyone to get an unbiased, and discipline-specific answer. Add to this that some LLMs can now capably read drawings and it’s easy to imagine how this would be useful for anyone on a project.
LLMs excel not only in translating between languages but also in bridging ideas, content or areas of expertise. In fact, the underlying technology is based on something called “transformers”, which are especially capable of reshaping language to make it more comprehensible and relatable to diverse audiences.
Such a capability seems tailor-made for construction, where project teams comprise specialists from various fields who often grapple with effective communication.
This isn’t a distant far future; this is an inevitable next step for project teams. For context:
- AI’s role in construction is already rapidly expanding, delivering significant gains.
- Many organisations have already started integrating some form of AI into their projects, using AI-powered tools for design, information management, report gener¬ation and communication.
- Although the rate of AI adoption is unpredictable, tools like ChatGPT and Midjourney have demonstrated how quickly AI can be adopted and add value.
How can LLMs and/or AI benefit the construction industry?
With the rate of change and progress in AI and LLMs, it’s guaranteed that our expectations will be exceeded, and that these predictions won’t age well – but I’ll attempt some anyway. Imagine:
- Generating executive summaries from any project data, tailored to your preferred style, area of focus and length.
- Rapid prototyping of design options with complete and accurate summation of cost, time and carbon totals.
- A real-time early warning system that pinpoints significant project delays and risks.
- Real-time analysis of team communication patterns, gauging project tone, health, morale and culture.
- Granular insights into project progress, with actionable improvement suggestions.
- Predictive tools that provide strategies and recovery plans based on the project status.
- A refined query engine that anyone on the team can use, for status updates, task support or information recall.
- The ability to scan across a whole project and be directed, and zoom in, to issues or key activities that need attention.
Many of us already use LLMs for more mundane tasks – from note taking to drafting reports. Given the high stakes, substantial investments and industry demand, exponential adoption is certain.
All of this will lead to the outcomes our industry needs. Our projects will be far more predictable, consistent and delivered with reduced risk and improved safety.
Are you ready for the future?
If you’re wondering what to do next, I’d recommend checking out some of the links and exploring these tools, especially ChatGPT. Engage with it, not simply as a better “search” tool, think of it as a conversation, and encourage it to be more active and specialised.
If this glimpse into the future interests you, join the conversation. I’m launching a workshop series on AI’s role in construction. For an invite, contact me using the details below. We’re looking for guests who are interested in this subject from a wide range of backgrounds and expertise.
James Bowles
Founder
Freeform
Tel: +44 (0)7989 232421
