Which Parts of Graphic Design Are AI Actually Repl

I’ve been thinking about how quickly AI design tools have improved over the last couple of years.

It’s easy to ask, “Will AI replace graphic designers?” but I’m starting to think that’s probably too broad a question.

Some tasks seem much easier to automate now—generating variations, removing backgrounds, creating rough concepts, resizing assets, or producing initial visual ideas. But other parts of design still seem to depend heavily on understanding the client, the audience, the brand, and the reason behind a particular design.

I’m curious how other designers here are seeing this in practice.

Are there particular tasks that you used to spend a lot of time on that AI has already changed significantly?

And do you think AI is mainly replacing certain production tasks, or is it eventually going to reduce the need for designers themselves?

I’m especially interested in hearing from people who have actually incorporated AI into their workflow rather than just opinions about AI in general.

Has it made you faster, changed the type of work you do, or created new problems that weren’t there before?

Read this on Internet : “Canadian employers are paying the price after AI proves unable to replace laid off staff”

That’s an interesting example. It shows that AI can automate parts of a job, but replacing the experience and judgment of a skilled professional is a much harder problem.

AI is not human and will never become one. Some people believe AI can replace humans, but that’s impossible. In fact, AI often makes more mistakes than humans because it just connects to a database for answers—it doesn’t think. AI can help to resolve some issues, and it can help to automatize some jobs at the moment (repetitive jobs).

I suspect the biggest change won’t come from AI automating existing tasks. Instead, the biggest impact might come from reorganizing and disrupting the entire workflow. Instead of speeding up the existing workflow, it will break it by creating better and more efficient pathways to the final product.

You asked for personally experienced examples, so here’s one. Recently, a book publisher contacted me about designing a series of romance book covers. A year or two ago, I would have turned down the job.

First, I’ve never read a romance book, have no interest in them, and would struggle to get into the heads of those who do. Second, the publisher was quite specific about the cover imagery, which would have required hiring an illustrator, a step the budget didn’t permit. Third, I like to read the manuscript before designing a book cover to make sure the cover accurately reflects the tone of the book, but again, the budget wasn’t enough to allow me to spend several hours reading it before starting.

Instead of turning it down, as I would have done two years ago, I accepted the job and relied on AI as a partner in designing it. I uploaded a tightly written creative brief to three AIs, along with a PDF of the entire manuscript, and asked them to read the book, compare the subject matter to the creative brief, and make written suggestions.

I selected what I thought was the best set of suggestions (Claude’s), then refined the written concept over the course of a few discussions with the AI.

I took that and created a series of prompts for another AI (ChatGPT) and asked it to generate the imagery in the style of several illustrations I thought captured the book’s mood. After several iterations, I had a perfectly usable illustration. I used Topaz’s upsampling AI to increase the resolution. I used Photoshop’s generative fill features to fill in the back cover and create something that matched the book’s aspect ratio. I even asked Google’s Gemini to write a blurb about the author for the back cover based on information it found about her on the internet. I moved it into InDesign to add the typography and other vector elements, and I was essentially done.

Instead of spending hours reading the book, looking for and hiring an illustrator, hoping for the best, making and refining a series of preliminary sketches, wasting time in back-and-forth discussions with the illustrator and the editor, and making who knows how many revisions, I started and finished the cover design in just a few hours, and the publisher accepted it with no changes.

Using AI as a partner in various aspects of the project was a completely different workflow than what was possible two or three years ago.

2 Likes

That’s a really interesting example, especially because AI didn’t actually replace the designer in your workflow it expanded what you could realistically take on.

The part that stands out to me is how you used different tools for different stages rather than expecting one AI to do everything. The creative judgment was still yours: deciding which suggestions worked, refining the concept, choosing the visual direction, and handling the final typography and layout.

I think that’s probably where the bigger shift is happening. Designers who learn how to integrate AI into their existing skills may be able to take on projects that previously weren’t practical within the available time or budget.

1 Like