AI as a tool for more complex designs

This is how it happens now for a lot of jobs. The client sends us the leaflet design and we print it. One new wrinkle though - the client sent us his design and asked us (me) to change it to incorporate products from a previous leaflet.

This was next to impossible to do the normal way (editing in PhotoShop) because of the complex background. We used AI to separate the items from the background and give me files for each item and the background. It took minutes to do what I would have taken days to do properly.

There was also a flash headline with a complex fill to the lettering, with a 3D effect and highlights. I could have done this with a few days, but here’s the important thing - I would never produce a design that complex because time and therefore costs would not allow it.

In general, AI leaflet designs I am seeing are not beyond my ability to create, but I would never contemplate such designs because of the time involved.

We are increasingly using AI to create impressive and complex designs in-house. Looking at the results next to my own efforts, I have to agree that the AI designs are superior. I still insist on doing the text separately because of the bitmapped nature of AI output.

The fact that almost anyone can produce better results using AI is making me think about retirement. I am learning to use AI and the changing workflow is easy to adapt to, but if a beginner can produce the same quality of design then what am I doing?

1 Like

I agree and disagree with parts of this.

I think AI is going to seriously affect the lower end of the design market, particularly the €5 business card or €20 poster jobs where clients expect the world for very little money.

And to be fair, I’m not particularly concerned about losing that market. I’ve largely moved away from those types of jobs anyway. Spending days on a design for someone who doesn’t value the time or expertise involved simply isn’t sustainable. I’d much rather concentrate on meaningful projects with reasonable budgets, whether that’s the design itself, print production, publishing, or a combination of all three.

If a local hairdresser, takeaway or small shop can produce a decent promotional poster using AI, fair enough. They probably weren’t going to commission a professional designer at a realistic rate in the first place.

Where I disagree is the idea that producing something visually impressive necessarily makes it better design.

AI can produce some genuinely impressive artwork, particularly complex backgrounds, illustrations, effects and imagery that would traditionally take hours or days. But complexity doesn’t automatically equal good design.

One thing I’ve noticed with a lot of AI-generated posters and promotional material is the lack of a clear visual hierarchy. There’s often so much happening, with so much information competing for attention, that the actual message gets lost. Instead of a poster communicating something in a few seconds, you end up with something that takes five minutes to read and decipher.

And that’s where design experience still matters. Knowing what to leave out can be just as important as knowing what to put in.

Then there’s the technical side of things.

It’s one thing producing an attractive image for a one-off promotion. It’s another maintaining consistent branding across dozens of products, publications, packaging formats and advertising campaigns.

Colour consistency, typography, legibility, accessibility, correct artwork dimensions, print specifications, separations and production requirements don’t suddenly disappear because AI generated the artwork.

In regulated industries, there’s an even greater responsibility to ensure that information is accurate, readable and compliant. Something can look fantastic on screen and still be completely unsuitable for its intended purpose.

And even outside regulated industries, will the AI-generated design reproduce correctly? Is the small black text made up of four colours? Are the colours consistent with the company’s existing branding? Does the artwork work at different sizes? Can it be edited and adapted six months down the line?

Some clients won’t care about these things, and that’s fine. But plenty of businesses absolutely do.

I actually think your approach is the sensible one. Use AI to do the heavy lifting where it makes sense, then apply your experience to turn that output into something technically correct, consistent and fit for purpose.

I don’t see much value in spending three days manually creating something that AI can produce in ten minutes, provided the result is suitable and the necessary checks are carried out.

The skill isn’t necessarily in how long it takes to produce something. It’s in understanding what needs to be produced, why it needs to work a certain way, and whether the finished result actually achieves its purpose.

AI is certainly changing the profession, and I don’t think we should underestimate how disruptive it will become. But I don’t think a beginner generating an impressive image is necessarily demonstrating the same skills as an experienced designer delivering a complete, reliable solution.

Perhaps the value of the designer is shifting away from simply creating the artwork towards directing, refining, validating and managing the entire process.

And personally, I’d rather embrace the tools that make the tedious parts of the job easier than spend my time competing with them.

1 Like

^ This.

Considero que la inteligencia artificial se ha convertido en una herramienta que está transformando significativamente la manera en que entendemos y ejercemos el diseño gráfico. Me parece interesante cómo, en el ejemplo, la IA permite resolver procesos que antes requerían días de trabajo, especialmente cuando se trata de separar elementos complejos, trabajar con fondos o generar efectos visuales que, por cuestiones de tiempo y presupuesto, quizá no se habrían considerado.

Sin embargo, creo que debemos distinguir entre crear una imagen visualmente atractiva y desarrollar una solución de diseño con intención y propósito. Aunque la IA pueda generar resultados que incluso superen técnicamente algunas propuestas realizadas por un diseñador, no necesariamente comprende el contexto, las necesidades del público, el mensaje que se quiere transmitir o la identidad que se busca construir.

Desde mi perspectiva, la IA no debería verse como una competencia que amenaza nuestra profesión, sino como una herramienta que nos invita a evolucionar, explorar nuevas posibilidades y optimizar nuestros procesos creativos. El verdadero valor del diseñador no está únicamente en saber ejecutar una técnica, sino en tener la capacidad de conceptualizar, tomar decisiones, comunicar ideas y darle sentido a cada elemento de una composición.

Me parece que el reto no es competir con lo que la IA puede producir, sino aprender a aportar aquello que todavĂ­a depende de nuestro criterio, sensibilidad y pensamiento creativo.

Same here, the AI part is the easy bit now. What still catches us out is what happens after, when the file has to separate properly. A bitmap background with a foil or spot colour plate on top, or small text built out of four colours, will look fine on screen and fall apart on press. So we end up spending the saved time on checks rather than on the artwork itself.