How to spot generative AI usage in book design: a guide for self-publishing authors and small presses
- whitestonepages
- Jul 23
- 5 min read

It's no secret that AI usage has been creeping its way into just-about every industry, and publishing is no different. Whether it's to tackle operational bottlenecks on a wider scale, or to save budget and outsource tasks on a specific project, evidence of AI usage can be found from the Big Five to the self-publishing sector.
As an independent author, small press, or self-publishing agency, you may be wondering: what are the tell-tale signs of generative AI usage? Whatever your stance on AI may be, I believe staying informed and AI-literate is key to navigating a changing industry landscape.
Here are some easy ways you can spot generative AI usage, from a professional book designer...
What is "generative AI"? 🤖
The official definition for "generative AI" is: "a subset of AI capable of generating text, images, video or other forms of output by using probabilistic models trained across [...] large amounts of specially-curated training data to discern and replicate complex patterns and structures. The output [...] mimics the characteristics learned from that training data, enabling many novel applications – ranging from personalised content generation, to advanced analysis and evaluation, and aiding creative processes."
Learning from a mass amount of stolen data and literary / artistic works is central to the controversy of using generative AI, particularly in a creative industry such as publishing where intellectual property is essentially the product, and copyright laws exist to protect them.
In the publishing industry specifically, usages of generative AI can include (but is not limited to) the following examples:
Using AI to create illustrations for the cover or interior pages
Using AI to write creative or marketing content
Using AI to produce marketing and promotional graphics or videos
How can we spot generative AI usage in book design?
1) The basics: illustration / design style and text errors 🚩
Let's start with the easiest signs to spot, which a lot of people may already be familiar with through their own day-to-day interactions with social media.
For illustrations: the flawlessly airbrushed finish, the extra finger or limb, the glassy-eyed expressions, the excessive use of light and shadow, and things that just don't make sense.
For text: beyond grammatical and spelling errors, sometimes it's in the wrong language, or appears in an unexpected place, or looks completed distorted or undecipherable like ancient glyphs.
Rule of thumb: if it doesn't look like it has been sense-checked by a human being, it was probably produced by a machine.
2) Odd positioning and object alignment 🚩
We book designers spend a lot of time thinking about how best to present text in a way that has visual impact and is easy to consume. AI-produced designs normally lack this intentional thoughtfulness.
You may find that your title and author name, for example, looks ever-so-slightly out of line with each other when they are both supposed to be centred on the page, or perhaps there are three or four different justification styles used to format one blurb.
If it looks like a regular book cover or interiors at first glance, but on closer inspection, feels more like someone threw text onto the page haphazardly without much thought, AI was probably involved.
3) Low resolution and the inability to increase it 🚩
Particularly at this stage of AI development, a quirky limitation I've noticed is that AI is not able to export designs to meet printing standards.
For example, it is well-known that commercial printing requires a minimum resolution of 300dpi. A professional book designer would ensure that all assets are created and exported in high resolution, and having the working files means that they can easily tweak export settings to meet printer requirements.
On the other hand, you can go back and forth with AI asking it to increase the resolution and it will confidently produce something that fails to meet the requirements. It will even hallucinate and tell you confidently that it is 300dpi, when, in actuality, it is significantly less than that.
(Tip: You can easily check a file's resolution on Windows by right clicking > show more options > properties > details)
4) Lack of contextual awareness 🚩
When you hire a book designer, you hire their technical expertise, practical experience, and knowledge of the market. When they receive your brief, they would usually start with understanding your book's genre, sub-genres, positioning and target audience. They would also complete market research into these areas so that any drafts are well-informed of current trends.
AI, on the other hand, works with whatever brief you provide them and draws knowledge from a wide database of published book covers from all different genres. It will fulfil your brief following your instructions, but may not necessarily understand the nuances of book positioning or how to get it right.
Book design is like a visual language that is spoken by designers and interpreted by readers. Is the book cover giving you thriller vibes when it's supposed to be a romance?
Is it supposed to be an adventure story but somehow reminds you of a business book?
If the cover is flawlessly executed but there's something about it that doesn't feel quite right, then AI might be at play.
5) Lack of evidence and layered files 🚩
As with any creative endeavour, it's a step-by-step process. When you are working with a human designer, you can expect to see evidence of creative work in the form of mood boards, rough sketches, first drafts, and, in some cases, even timelapses. Some designers may voluntarily share every stage of their design thinking with you and ask for your feedback, others might choose to work more independently. If you are working with a human designer, these receipts will exist and most would be happy to share them upon request. If you are working with a completed design or illustration, you should still be able to see evidence of layered files.
With AI, creation is pretty much instantaneous. In goes the prompt, out comes what could pass as a perfectly-executed final product. There are no stages in between (for those not privy to the back and forth of design prompts.) So if you ask a designer to see their rough sketches, first drafts, or layered files, and they are being extremely evasive about it, the likelihood is that there are simply no receipts to share.
Conclusion
At this moment in time, with AI at its current stage of development, these are the top five signs that sends my AI-senses tingling as a professional book designer. The more of these you can spot in one piece of work, the higher the likelihood that generative AI is at play.
That being said, I feel the need to put a massive disclaimer that human errors do occur, and an occasional poor alignment or design choice doesn't necessarily mean that your designer is a bot. Ironically (to this post), mistakes are part of being human!
It will be interesting to revisit this blog post in a few years time to see whether these signs still stand the test of time, or whether AI will have overcome these limitations through further advancements.
Until then, I will continue to observe and share my learnings!

Hi! My name is Wendy. Based in the UK, I'm a book designer and typesetter with a Master's degree in Publishing and over 6 years of industry experience. I love helping indie publishers and authors bring to life their manuscripts through great cover and interior design. As a designer, I am curious, communicative, and continuously learning. When I’m not reading or designing books, you can catch me cosy-gaming, journaling, and obsessing over self-development!


Comments