How design is embracing the AI challenge

A little over two years ago, Liquid's nascent Future Led series tackled super-intelligent AI and we were told that the pace of technological change would be “very dramatic”.

Just how dramatic and world-consuming that change would be - and continues to be - is mind-boggling to consider now. So, we returned to the topic to re-engage with what's happening, how it's reshaping the landscape of design, and how other industries and disciplines are embracing the AI challenge.

Joining our panel this time was:

  • Shae Quabba, Design Anthropologist & Experience Innovation Specialist at Suncorp
  • Anne Kovachevich, Sustainability Buildings & Precincts Lead at Mott MacDonald
  • Michael Collett, Conversation Designer at the ABC
  • Sam Daley, Liquid's Head of Product

While AI has the potential to enhance efficiency and expand creative boundaries, the challenge lies in striking the delicate balance between automation and empathy.

All the speakers were "techno optimists", choosing to be excited about AI's future use cases. But, of course, they urged caution amid the very real challenges that exist in the friction space between human experience and algorithmic precision.


1714x2284-website story FLS 3.3 smaller file

(L-R Shae Quabba, Liquid's Sam Daley, Anne Kovachevich, Michael Collett, Andrew Duval)


AI as a tool

As a design anthropologist, Shae combines social sciences and design to understand how humans do things. She told us that, for her, the intersection of digital experiences and AI was fascinating.

"The biggest change that I've seen is with us as designers and how we use tools," she said. 

"So using Midjourney to create personas, using different AI elements with the tools we use like Figma or Adobe, [and] tapping into large language models to pull together content," she said.

"The ultimate challenge is how might AI augment the way that we garner insight? Because that synthesis piece is really fascinating to me. How smart is an AI to be able to derive the same kind of insight and result as a human?"

She added that improving tools for customers was also an important part of the equation.

"We've been working with various aspects of AI for a number of years, that consumers probably wouldn't really know that they're interacting with as they go through the experience," she said.

"For example, we've been making claims processing faster for a number of years. You essentially give us [some information] and classifiers are working in the background to make sure that your claim is processed faster … making the experience faster."

For Anne, whose work focuses on sustainability, using AI to optimise outcomes is a huge positive.

"We've had some of these tools for a long period of time, probably for at least five years, and we were initially worried that Google would try to take over all of the engineering jobs, and they would try to design all our buildings with algorithms.”

"I think what we've found is that we actually need a sanity check. We need someone who knows the craft of engineering and how to design different systems, and be looking at the results and teaching these systems how to optimise."

A steep challenge in Anne's industry has been moving towards sustainable thinking and building. For this reason, she believes AI can also contribute to changing mindsets. An example is being able to run simulations to explore future scenarios and make sustainability decisions sooner.

"We're always trying to get in really early when our architect might have some kind of concept model. What's really advanced these days is that we can get a very simple model, and put it into a simulation where we're looking at the local climate.

"What are the different climate responses that we need to think about? We not only look at what's happening now, but we look at the future, because in 50 years' time all our buildings and precincts are still going to be existing and working, so we need to make sure that we're thinking about what that future climate is and make sure that buildings are designed for that.

"So we might do a simulation that says, okay, this is what the climate is now, this is what it would be in the future. Then we might run that model to look at the daylight hours and optimise different shading components. Or we might look at the systems within that, so what kind of air conditioning systems or where can we get rid of air-conditioning systems?"


But it's like 'trying to control a wild horse'

Having a new tool is great, but as Sam explained - it's currently still a blunt tool.

"I think it is partially knowing what it's good for and what it's not. It's a blunt instrument that's controlled by blunt instruments," he said.

"If you've got fairly loose expectations of what you need out of it, it can absolutely exceed those expectations, but the more you need a really precise requirement out of it, suddenly it gets harder and harder to control.

"It's a bit like trying to control a wild horse. It's powerful, it's fast, and you can kind of get under control to a point, but if the requirement came through that you need to have this 99 per cent degree of precision - which is like trying to steer a horse through a narrow hallway at speed - well actually, we don't know if can do that."

This was echoed by Michael, who can see the huge potential to use AI with his work, but the existence of even a small risk makes him pause.  

"I think my biggest frustration is that the most exciting ideas that I have for how AI could be used do actually require 100 per cent precision. And that's just not there," he said.

"Because I need to be pragmatic and think about things that we can actually achieve now. And some of those are a bit behind the scenes and might seem a little bit more boring than what we had been talking about with generative AI, but there's some really awesome use cases there - I'm thinking in particular about things like search and stuff like that."


The trust factor

That lack of precision means things like trust start to become part of the conversation.

"The challenge of designing for trust is a fabulous challenge, and I'm quite optimistic about it, even though there are huge downsides," Shae told us.

"If you think about what you do as a designer, you really want to untangle that ball of wool that the cat has got into. And I feel like that's kind of what's presented to us at the moment - designing for trust and how does that trust come across in brands?

"How do we use lots of different tools as designers to make sure that that principle of trust, that that design guideline is always there in everything that we do."

She added that this was particularly relevant for customer-facing brands or services. People were starting to understand what AI can and can't do, but Shae said customer behaviour was still showing reticence to engage in some cases.

"I think there's a trust hurdle we've got to get over," she said. "[We keep] crashing up against [customer] expectations.

"We will take content out of a chatbot design and we will put it in front of customers in another way, and they'll say, yeah, this is great. 

"The frustration that we have is the legacy of the past experience that sits in people's minds, their mental model. People will not use chatbots because they think, oh, this thing is dumb."

For Michael, that trust factor is, as you'd expect, a priority for institutions like the ABC.

"That's the mystery of these things and that's my frustration; you have these ideas and you think of all the pros and then you get to the cons and you're like, oh yeah, that's right. We can't trust this thing. 

"I think what scares me specifically coming from an ABC context is that we are not going to be able to play that same game as aggregators like Google and Bing. For us, trust really is paramount. So does that leave us behind in those sorts of areas? I don't know."


Powering more human experiences

Despite a lack of precision and the trust deficit right now, Sam said he believed AI and its capabilities could actually help make more human experiences.

"Prior to the opportunities created by large language models and things like that, I've often felt that our digital experiences aren't human enough as they are now," he said.

"Quite often the technical landscape and the kind of experiences built on these commercial off-the-shelf systems … are efficient, but they're often containers for experiences that aren't very personalised.

"They're containers for content and they're good at making facts and information available, not necessarily for providing guidance or for providing support in a way that feels human.

"From a business point of view, that makes a lot of sense, but there's often a bit of a lack there. 

"I think that's where Liquid's strength has always been: how do we bring that human experience in there? How do we build an experience layer over the top of that distributed, federated landscape of technology that helps guide people through, that helps create experiences that feel fit for purpose?

"That comes from human-centred design, understanding what people need and want from an experience, and then building and crafting those kinds of experiences."

In Anne's field, the human experience is living and working in a sustainable way.

"We are very much looking at how we move from a linear economy to a circular economy," she said.

"So the ability to really understand all the different components in our cities and how we can potentially reuse or repurpose those? How can we really look at how we design things so that we're using less stuff and appreciating what's there. I think it's really exciting. 

"I think enabling the circular economy and creating data banks out of our buildings is super exciting. And that's all enabled [by AI]."

More from our Future Led series