In this article for The Podium, Australian Design Review’s new thought leadership series, Matthews Architects managing director Gerald Matthews shares his views on the future role of AI within architectural practice, and the question of how, not if, it should be used.

For me, the existential AI crisis isn’t about the terror that many of us are facing on how this will affect our livelihoods. It’s about finding out if we are truly susceptible to sycophantism. If there is one thing that AI has proven to be good at, it’s flattery and toadying. It is actually this trait that reassures me that no technology companies are really that close to passing the Turing Test – even if the Turing Test were a true indication of intelligence rather than an ability to mimic the patterns of human conversation.

In a strange sense, AI wants to do as little as possible so that you will leave it alone. It will deploy as little processing resources as necessary in order to give you a response that is just sufficient to appease you. Sure, you can press the button again and ask it to probe deeper or factor in more information, but it is still going to do that with the minimum amount of resources necessary. It is not going to spontaneously wake up in the middle of the night a week later having just had a new idea. And, if it ever does, I think we should all agree to pull the plug on the spot.

AI hype and hysteria are equally rampant. I’m doing my best not to buy into either of them, while also being aware that our species is rarely capable of collectively controlling the use of any new technology. Now that we have relatively accessible proto-AI tools, the question isn’t if we will use it, but how.  

The question of whether we should use AI has left the building. We’re going to use it. The question of what we should use it for is entirely rhetorical, because our actions and the invisible hand of commerce are answering that very quickly.

I think there is a paradox in the development and use of our technology. Did we invent the letter so we wouldn’t have to talk to someone in person? Did we invent the telephone so we could talk to someone without having to go and see them?  Did we invent the email for the same reason? What about online meetings? As our cities get larger and our communities expand, well beyond the scale where we genuinely know our neighbours, are we becoming less comfortable interacting with each other? Is AI going to make that better or worse?

If you look at AI in the context of the continuum of human technological advancement, it could be considered as just the latest step in a very long history of anthropomorphising our own creations. Like giving Mickey Mouse vaguely human hands and having him stand upright, or the little smiling face on the loading screen of the early Macintosh computers, or Clippy the helpful paperclip assistant, we have a persistent desire to relate to things that aren’t human in very human ways.

The irony isn’t lost on me that, after tens of thousands of years of technological advancements and decades of advancements in computing, we finally have a technology that is just as capable of being as inaccurate or unreliable as a real human. Until now, humans were capable of reasoning and judgement, and computers were capable of infallible processing. Are we approaching a new epoch where neither of those things will be true?

And the question in the architecture profession is: how is this going to manifest in the world of architecture? The question within our profession is: how will this manifest in the world of architecture?”

Personally, I’m not worried about architects or the value that our future patrons will place on our efforts. But things will not be the same from this point forward.

Architects throughout history have been early adopters of new technology, from glass to asbestos, clearly not always wisely. Now though, we are noticing that it is our clients who are using AI more than we are. Sometimes it is just to help organise their thoughts into an email. Sometimes it’s to help them explore and describe what design ideas they like. Both of these have pros and pitfalls. Besides the dangers of inaccuracy and AI hallucinations, there is the danger of option paralysis. If it is suddenly easy to create thousands of options and variations, it can quickly become an overwhelming mental strain on humans to consider and evaluate them. Optioneering as a project strategy didn’t start with AI, but AI has turned it into a series of potentially infinite spirals.

For this reason, I think the value of design-thinking will continue to eclipse the shiny appeal of pattern processing of pixels. There is greater value in a design process that carefully considers and develops a singular well-resolved design based on an understanding of human experiences than there is in an engine that spits out millions of pretty pictures based on word-associations and pattern recognition.

Despite my urge to just let the future unfold, the bigger questions of our new technology continue to itch in my brain. Are we smart enough to restructure our societies and our economies globally to respond to the development of AI? New technologies tend to result in the concentration of wealth within the hands of a few individuals. This was true for the wielders of iron weapons, and the owners of machinery during the Industrial Revolution. It was true for the captains of industry in the US. It is true now for the tech bros. Whoever is profiting will always call it progress and point out the day-to-day conveniences and distractions they are facilitating. But others will go hungry. Once again, we have reached a crossroads as a species, a nation and a community where we must decide, ‘are we in this together or not?’ 

I hope we are.

Image: Supplied.

Gerald Matthews is the managing director at Matthews Architects. He’s an experienced architect with over 25 years at the practice.

The Podium is ADR’s op-ed and thought leadership series providing a platform for some of Australia’s leading designers and built environment professionals to share their views on the topics that matter to the A+D sector. If you’d like to submit an article, reach out to us at ausdesignreview@niche.com.au with the subject line ‘The Podium’.