FutureProof: How are architects using Gen AI and what are the implications of its use in the future?

This article explores how Gen AI is currently being used by those in the world of architecture.   

How are architects using AI?

BIM

Architects are already ahead of the curve in terms of using modelling systems and software; for example, BIM (Building Information Modelling) Technology has been around for decades and is used to support in the design, planning and construction of building projects. However, the development of Gen AI is a significant opportunity to streamline the design process, and provides new opportunities. For example, AI can be used with BIM to flag potential problems with the design.  

Nearly half the profession using AI already

RIBA’s AI Report 2024 found that around 50% of architects say that they have a basic knowledge of AI and 41% of practices have already started to use AI in some way. The way in which architects are already using AI seems to vary from using ChatGPT for quick searches and drafting emails, to creative uses in concept ideation, data analysis, the preparation of contracts and even using on-site facial recognition software for security on construction sites.  

Anecdotally, it seems that smaller practices are less likely to be adopting AI, or if they are, it is more likely that they are using the more widely available software (like ChatGPT and MidJourney). However, bigger practices with larger R&D budgets are making significant moves in developing their own AI which could give them an edge over their competitors. 

What are the benefits of using AI?

The potential benefits to the profession could be significant. 

Efficiency and cost

The use of AI to develop large design resources could help to increase efficiency and reduce cost, and notably, time in the design of new buildings. Use of AI modelling could also help to combat bigger challenges; for example, by giving architects the ability to analyse a variety of different designs in shorter timescales, AI could help create lower carbon and safer buildings. The use of Gen AI to enhance a process known as digital twinning is helping architects to simulate and optimise a building’s performance, including its environmental impact, helping architects to deliver greener and more sustainable buildings.

"Back office" work

From a “back office” point of view, AI can be used to re-word reports, prepare client presentations, or for the preparation of contracts. Where an architect is appointed as a contract administrator, the use of AI can help with everything from project management through to cost tracking, document management and risk management. Each of these uses has the benefit of freeing up the architect to spend more time on the creative and technical parts of their role.

Design creativity and optimisation 

Gen AI tools can be used to generate hugely creative design concepts whilst also taking into account criteria such as space requirements and design constraints, helping to speed up the design phase of a project. Effective use of AI can also assist with generating and reviewing/producing design variations quickly to assist with, and add value to, the design process. 

What are the risks?

Risk to jobs?

A third of RIBA’s respondents consider AI poses a threat to the profession. However, for now at least, the prospects of the technology becoming sophisticated enough to put architects out of a job is minimal.

When AI goes wrong

As with the use of AI in other professions, the same risks apply in terms of the possibility of bias or hallucinations (ie, the AI gets things wrong). For a construction project, if those issues are not spotted early, the ramifications could be serious - errors could result in a building that is over designed (which results in wasted costs) or, in the worst case, a building that is unsafe.  

This means that even if the software can be used as a tool to assist in the design of a building, the architect’s experience and expertise will still be required to ensure that the output is safe and workable. There are complex issues that go into a construction project including regulatory issues, aesthetics, and the relationship between various stakeholders, which mean that making a model work in the real world will still require specialist human input. 

For bigger practices, designing their own tech represents an exciting and potentially lucrative opportunity to streamline their practices; however, those organisations need to consider the risks if those systems go wrong and ensure appropriate safeguards are in place to support that technology.    

For smaller practices, using programmes like ChatGPT requires a cautious approach because of risks to data security; these models use the data that is inputted to them to learn, so using confidential data in an open system should be avoided. 

From a liability perspective, what happens if the AI fails? The architect will still have liability to its end client for their design without necessarily being able to pass that liability through to the AI provider in the way they may currently be able to do with a subconsultant.

Insurance policy coverage

Linked to the above, careful consideration will need to be given to the architect’s PI position. Will cover extend to AI related design issues? Insurance cover, and the information provided to insurers around the use of AI and cover for AI related errors will need careful consideration. 

Our content explained