Issue 02 , 2024 - AI in Design

Welcome to our second 2024 edition of RADAR – our monthly email. A 4-minute read from a design viewpoint to expand your perspective and deepen your knowledge, helping you better navigate the shifting world of brands and consumers.?

This month, we are exploring the Role of AI in Design from a designer's perspective, with thoughts from our Creative Director Andrew Capper.

Email [email protected] to receive them direct to your inbox.

The Role of AI in Design

Over the last 3-4 months I’ve immersed myself in Discord/Midjourney, Krea and Vizcom with a desire to better understand the capabilities of AI tools in the design space, and if and how we integrate it into workflow in ways that benefit our clients and creative teams.

AI in commercial design raises lots of questions. Should I/my agency be using it? When should it be used? Will it save time and money? How valuable is the output and what should I use it for? Are there risks?

All these interfaces have different strengths and ways of approaching creative generation, but what they all share is speed to generate convincing imagery and seemingly endless nuanced conceptual options.

I think the first thing to point out is different design fields will have radically different views and uses for it. I write with 28 years in branding, structural packaging and sustainability, designing better functional and emotional experiences for some of the world’s most loved brands, decoding values and fusing those into unique physical form. By contrast, for an illustrator creating other-worldly visions for a communications campaign, it may offer different benefits. For a print designer creating an autumn apparel collection it may deliver others. Every creative field will have a different relationship with AI and unique needs, uses and objectives.

?

Is there a defined brief?

What’s really apparent with AI is you get out what you put in, and the 2 key parts to that equation are ‘the brief’ and the ‘design team’.

If I Google ‘restaurants’ without defining a location, budget, ambiance or cuisine I’ll get back an infinite list of places to dine and almost all will fail to meet my needs. It’s much the same with AI.

AI is not a tool to plug the absence of a clear brief or business objective. Without key values or visual reference, and a sketch vision of the right direction to take, what’s generated is scatter-gun with huge amounts of time spent going down rabbit holes.

When is AI a Valuable Tool?

Firstly, its ability to quickly lift those first rough, back-of-an-envelope thoughts into something client-friendly and help filter out themes or territories that don’t show promise. Getting a feel for poise, volume and presence quickly is a joy that other software can’t match.?

Secondly, it’s incredibly useful when innovating and there are many thoughts and fledgling opportunities to bring to life and explore. Normally it’s uneconomical to manifest 50 raw ideas and a sketch can be easily dismissed. AI changes all that. As a tool to capture many visions quickly and convince the less design literate that an idea has merit, it’s invaluable.

Finally, whilst it can’t yet match the ultimate output of a high-end visualiser, at an early stage where you’re okay with less nuance and control of every single aspect of a design, image style and composition, it’s a powerful tool to bring a brand world to life and take stakeholders with you.

The watch-out to all these points is the need for clear communication of what AI generates and the need to manage expectations. We’re all on a journey of ‘slicker, sooner’, ever since Macs and Silicon Graphics machines entered the studio. I’m as big a sucker for a sexy visual as the next person but, as a designer, I can peel away the gloss to judge the merit of the idea beneath that the untrained eye cannot. The higher the quality of visualisation earlier in the process, the greater the assumption that it’s 1) a great idea and 2) a fully considered design, which can result in risk and disappointment.

When isn't it a good idea?

In the same way I’ve moved beyond a sketch pad by the time I’m refining a design for manufacture, I won’t yet be using AI beyond initial broad ideation. Whilst there are lots of different interfaces, none offer the control and nuance I need to craft forms to an appropriate level of integrity and accuracy. I don’t believe there are any shortcuts here and the need to see things in 3D and make a prototype to hold in my hand is a vital part of the process.

The questions around ownership, IP and confidentiality aren’t going away either. I’m comfortable using it to help focus down to a general concept direction, but not a defined solution. I already know of one large agency group that has banned the use of all AI tools until the legal minefield has been tested. Imagine a scenario where Brand A sues Brand B for copyright infringement because an aspect of their brand or packaging is deemed too similar. I’d be sweating if my defence was, ‘We put a sketch alongside some found imagery of other packs into an AI tool and Bob’s your uncle’.

Quantity vs. quality vs. speed?

One of AI’s strengths is the sheer speed to generate many alternative convincing ‘concept visions’. In a few hours I can have many, many design themes and iterations on each. Iteration management becomes a key skill to learn as a lot of them are just plain ugly or off brief.

I can still generate a dozen concepts, covering different aspects of the brief, faster than AI – they just won’t look as polished.

Do marketers feel they are lacking breadth of choice, or do they want a set of tight curated solutions that are all a right answer?

?

And if the output is going to need to be deeply interrogated by stakeholders and consumers, does AI give the answer to ‘why’?

?

Much of what I do is to connect business with brand strategy via design to explain why a form or experience is right for a brand and meets the brief. This generates belief and trust in an idea.

?

A race to ubiquity?

One of the main benefits is also an Achilles' heel, particularly using text prompts. Put in ‘carbon fibre bicycle’ and you’ll get exactly that aggressive looking Olympic cyclist’s machine complete with moody lighting that I bet you were all imagining. That’s how carbon fibre bike companies generally present their bikes to you, which in turn makes up the most numerous visual expressions in the data set AI is drawing from.

?

So that may help you develop something that fits the category, but it’s not going to help deliver uniqueness.

?

The Future

Well, it certainly isn’t AI in the way Arthur C. Clarke envisaged it yet. The usefulness of output today is still a direct reflection of the user’s input ability - imagination, understanding of the brief, creativity and technical skill.

?

That could equally be a description of Photoshop, a pencil and paper or a lump of clay. This would suggest that AI visualisation is another tool to add to the designer’s toolbox, not a magic bullet.

?

And now everyone has access to it, developing similar skillsets, does it become table stakes? In a commercial world where brands are looking for uniqueness and competitive advantage, the difference comes down to people, processes, experience and a bespoke approach to every challenge.

?

First defining what to design and then creating what has yet to be, rather than synthesising what has come before.