Art in the Age of Generative AI: Illustrating the ‘Engines of Engagement’

For a book about Generative AI , a book that explores the impacts of the new Art Engines on creativity itself, we wanted to do something special with the illustrations. This desire led to a collaboration between myself and both Sae Schatz and Geoff Stead , my co-authors (and now co-illustrators!).

Whilst i guess that people engage with my illustrations online, i do not describe myself as a digital artist per se, in that whilst i create the art on my ipad, i do so in very analogue ways. I use the tool Paper, by WeTransfer, that is really little more than a sheet of paper. There are no fancy tools or layers, just a simple palette of pens and brushes and a limited space to work in. This suits me very well.

Whilst Sae is also a very talented artist, she also has had formal training in graphic design, and used more sophisticated tools in the Adobe suite in her work.

Geoff is more like myself, in being an enthusiastic amateur, although a much more capable technologist.

For this book, we developed an iterative process, using new technologies, but which i think remains an essentially human creative work.

After discussing the ideas in the book, both Sae and Geoff went off and used Adobe Firefly (one of the new Generative AI tools, one of the new Art Engines) to create a folder of images. In doing so, they also used some of my original artwork to flavour the process (Firefly can use reference images to calibrate outputs).

We then reviewed these images, and i’d choose one to act as inspiration for my own original work – or sometimes i’d use elements from several of them – or just make up new parts in my head as well. In this sense, the Art Engines (and the human expertise behind them) primarily gave us a composition, context, and individual elements. It also created a visual style that i felt was recognisable as mine – that felt comfortable – but that also carried me into a new style and space.

There were some very specific ways that i was able to travel further in my art through this process: there are i think six illustrations of faces in the book, something i’ve never previously done, as i am unable to draw faces (this was one of the very first things i wrote about on the blog, well over a decade ago!).

The resulting images i then shared back to the group, anxiously awaiting feedback! When we had something that felt right, Sae would work on the new image further (and Geoff would sometimes use it to revisit or create new foundational material).

Through trial and error more so than by design, we settled into a certain palette of colours and more importantly a visual reference and style which for want of a better term i think Sae described as ‘fractured’!

Sae also found certain common approaches to enhance the images: often pulling out elements to give them a shadow and three dimensional quality, to pull fragments out of the page, or occasionally to use the infill capability of the Art Engines to add texture or fill in detail to a block or shape. She also used her more traditional media based skills to lay on certain twirls, patterns, and texture.

For me, the final images are a little ‘deeper’, through the texture, and often more vibrant in the colours. I’m also pleased with the visual language we have created.

There are imperfections in this work: inevitably, because of our iterative approach (and my somewhat limited creative ability) some of the earliest images (which we had been congratulating ourselves on creating) suddenly felt a bit left out and left behind. Our style had emerged and moved on.

Some we reworked, a couple we quietly executed out back. But an observant reader will be able to spot two or three that i hope they like, but kind of feel like outliers. I’m ok with that: the imperfection in my own work is part of the creative process. I would rather have work with integrity and brush strokes than something perfect but sterile.

I suspect if we had carried on longer with this process we would have continued to evolve: perhaps that will carry us into a second edition in time.. or there is always the Learning Science book that we are working on to try new things!

I have found the experience of using these new Art Engines for creativity very different from simply writing about using them, and i thought i’d try to capture my thoughts why.

Firstly, a tool like Firefly perfectly illustrates a core paradigm of many Generative AI tools: there is almost no barrier to entry, in that they operate within our existing dialogic and verbal conception. You literally say what you want. There is no coding, no particular learning curve. You can get from ‘nothing’ to ‘great’ in about sixty seconds, simply by saying ‘draw me a dog on a mountain staring into the heart of the universe’. A better picture than most of us could create with pens and paint.

The technology democratises that creativity, to a radical extent.

Almost immediately following that i find myself with a sense of dread, paralleled with excitement. Dread, because it’s immediately clear that much of the bread and butter work of illustrators and artists will disappear – not all of it by any means, but a valuable layer of the commercial ecosystem has had the bottom ripped out of it. Many of my friends who are artists do some of their earliest commercial work creating album covers for friends of friends, or posters and fliers for local events. Much of that will be eroded. Alongside that, much work that previously would have been text only will now carry world class illustration, which may be a good thing, in terms of the richness of our lives. But not the richness of our artists.

Creatively, the thing i have enjoyed most is the dynamic ability to play with composition, to create a scaffolding for an image. But in parallel with that i’ve felt outclassed at time, outmatched by the speed and vibrancy of the Engines.

I’ve also been surprised by the effectiveness of the referencing ability: i fed one of the Engines some illustrations i’d done for a children’s book, and with very little effort was able to create a sequence of wonderful images, that felt like ‘mine’.

But they’re not. I’ll never use them for my children. I don’t say that as a dramatic statement, just a very personal one.

I’ve always illustrated my own work, and the act of illustration is not simply about adding pictures to existing text: it’s often part of the creative process. Illustration – the act, not the artefact – is part of my language.

I often draw the illustration for the blog before i write anything, that act of illustration helping me to literally shape my thoughts.

I have no doubt that i could create images for my work far more quickly and easily than i draw them. Some days i spend fifteen minutes on an illustration, others it may be an hour or two. But they are part of the work: if i shortcut the process, i lose the benefit.

Scarcity itself can create value, and perhaps that’s part of my thinking. I probably create five new images a week, and have thousands in a folder on my iPad, representing over a decade of work. The other day i saw someone share a very early illustration in a PPT they were presenting. To me it looked awful, but clearly for them it illustrated something they wanted to say. My work has evolved, and continues to evolve: whilst i’m sure i will continue to illustrate my work by hand, i also have no doubt my work will be influenced further through collaboration or through new tools. Art is a process with no end. But sometimes finding simplicity can be key.

For ‘The Humble Leader’, the book i published last year as a guided reflection on humility in our leadership, i wanted to strip back my own visual style. I ended up hand drawing a series of twelve leaves, each with a mapping pen on cartridge paper, subsequently photographed for the book. I think there’s an honesty in that work that i simply do not feel in the digital, let alone the outsourced to AI.

But who knows: art is fickle, and should bring us challenge, provocation, and joy.