Human Software is a book about the importance of people working together to overcome the machinations of evil corporations. It’s a story about how humans are blamed for mistakes that computers make. A human story set in the modern world of software development, where engineers and managers are often pulled in competing directions simultaneously.
For such a human-centric, anti-AI book, why did I use AI to assist with the cover design and the internal illustration of the map?
This is the story of both the “why” and the “how” of the decision to use it, plus a look towards the future.
Generating the Idea
Two years ago, when I started writing Human Software, I had a clear vision for the aesthetic. It was crucial for the overall feel of the book and the appearance of the final product. I knew I wanted a boat, Beth and Dominic’s “Buffy”, on the cover. I knew I wanted the bridge, a cooling tower. An industrial feel with the freedom of the sea in the foreground. Many older books inspired me, as did many artists and stories I’d read. Iain Banks and his fiction novels of the 80s, including “The Wasp Factory”, “The Bridge” and “Complicity”. Alisadair Gray’s “Lanark” and his artwork. I also drew inspiration from artists such as Escher, Whistler, Edward Hopper as well as modern artists such as Kay A Brown – an artist from the South Coast of the UK whose specialises in wood block art.
The book is set in Kent, UK. It’s my home county in my home country. However, for dramatic reasons, I wanted to set it in a fictional place where I could have some fun. So, I created Sandport as an amalgam of the town of Sandwich and the ‘area’ of Richborough and Stonar industrial estates. Sandport was to stand on the Wantsum Channel – a long-gone channel that divided Thanet from the rest of Kent.
So I started to reimagine this part of the world, and then I redrew the map of Kent with the help of some research. I ended up with a hand drawn map.
The Initial Designs
Returning to the cover, I used the generative AI plugins in Canva to quickly generate some early ideas. The prompts varied from black and white focussed to Edward Hopper or Whistler-inspired covers. And while some of these were excellent, I eventually settled on the bottom-right version as a placeholder cover for the duration of the book’s completion.
Working with Humans
What I really wanted, though, was a hand-drawn, human-created cover and map. While I reached out to various individuals, I couldn’t find a person to help me match my vision to reality.
I’d already used the freelancer platform Reedsy to find an editor for Human Software. It also provides a platform for illustrators and cover artists. So, I posted my project and interviewed some potential collaborators.
Some of the reference images I included in my pitch were of the original black-and-white covers to Iain Banks’ books, created by the British artist Peter Brown. Notably, The Bridge. This book had much the same effect on me when I read it as did Lanark on Banks. I’m fascinated by the huge structures that we build – these bridges, these towers, these churches to industry. I wanted that feel to come through in the cover and the map.
Sadly, none of the people I attempted to work with could understand where I was going. So, I set the work aside for another day and continued editing and revising the book.
The Cover: Working version between May and August 2025
While I continued to look for potential collaborators, I continued using Canva to refine the cover. While I liked the latest version a lot, the use of Generative AI art for the whole background image really bothered me. I wouldn’t be publishing the book with AI generated art – that was certain.
I would have to find an alternative way of creating it.
Drawing a Conclusion
I knew that I wanted a bleak industrial landscape. I wanted to include sailing, and I envisioned a boat featuring prominently on the cover. During the course of writing the book, I decided that “Buffy” needed to be a Westerly Centaur 26. I wanted a hand-drawn cover, and I already had all the elements and the inspiration. So I ordered a new set of pens and set about drawing my own cover.
A boat. A distant shore. The cooling towers of the long-gone Richbough power station. The wharves of Sandport.
I’m very pleased with the result – not perfect, but not bad.
However, I wasn’t totally happy with the execution or the contrast in the original image I created on the left…. so on a whim I decided to ask ChatGPT to help me. I provided a scan of my source image and used a single prompt.
“turn this image into a black and white woodcut style with more definition and higher contrat (sic)”
ChatGPT thought for a few minutes and, in one take, did a stunning job of converting the image to a woodcut type. Before I could use it, though, I needed to print it out, make some manual adjustments to the image with my pen, scan it back in and then touch it up in Acorn (an excellent Photoshop alternative). Then it was in the correct resolution and clean enough for me to use.
Mapping the Future
Then I tried the same process for the map. Despite having already completed a detailed map, the results from ChatGPT were confused – the details caused glitches in the output. So, instead, I spent some time drawing a simplified version of the map without labels and then asked the AI to do its thing again.
The result was very dark but pretty impressive. It did miss assign the bottom right section of land to ‘sand’, so I had once again to make manual adjustments through editing, printing, drawing and scanning.
Once again I needed to modify the output from ChatGPT, print out the result, make some pen and ink modifications and then rescan the image. Finally I created some name plates (again with ChatGPT) and printed those out before placing them on the image. The resulting image was then scanned in to create the final result.
The Final Result
I moved the image of “Buffy” back into Canva to typeset the cover.
The map was scanned in – giving it a fuzzy look which I find quite effective as if one of the characters in the book (perhaps the security guard from the cooling towers, perhaps the historian) have created it themselves.
Publication and Beyond
When you self-publish your book through most platforms, in my case Leanpub and Amazon KDP, you need to declare if AI has been used in the creation of any of it, including images. When I publish, I will admit to using ChatGPT for helping with the creation of these images; however, I feel I have limited the use of Generative AI as much as possible. Creatively, I’ve had a vision, and I’ve used GenAI as a tool to help me achieve this.
I distinguish the assistance the tool has provided in my use case by simply using a prompt to generate the whole image. However, I still feel that even this limited use of GenAI is stealing from other professionals. The alternative, however, and one I considered, would have been to spend some time learning to make woodcuts. I’ve certainly not ruled this out for the future; however, I feel my expertise lies in setting the agenda, rather than executing it!
While the process I’ve followed somewhat alleviates the guilt I feel over this AI hypocrisy, I hope to be able to commission the work that will convert these images into proper woodcuts from a skilled artist in the future. As an independent, first-time author, you want to be able to do the best job you can with the resources you have at your disposal. I hope this post is helpful for other authors and artists who are likewise caught in a quandary.
I hope you will be able to enjoy the artworks in person when the book is published shortly on LeanPub and Amazon.