Published
13 November, 2024
by
Param Pratap Walia
Future-Proof Design: Adapting Creativity for an AI-Driven World
This project explores how designers can adapt to the rise of generative AI by incorporating semi-procedural workflows into graphic design. Inspired by the efficiency of 3D art pipelines, the project tests this approach using Blender to create algorithms that automate repetitive tasks, freeing designers to focus on creative decision-making. A case study on "Found Shapes" demonstrated that the semi-procedural method is 4 times faster and 500 times more productive than traditional design processes. This shift redefines the designer's role, emphasizing creativity over labor-intensive tasks and enhancing versatility in various design contexts.
Introduction
This personal project is a response to concerns about how the recent rise of generative AI services could affect graphic design jobs. This is a scary idea as I am preparing to enter the workforce and I want to explore how I can adapt to this change.
Historical context of technology’s influence on the job landscape:
Technological breakthroughs have reshaped the job market by both eliminating and creating job roles. These innovations shift how work is divided between the tools we use (known as capital) and human workers. According to Acemoglu and Restrepo’s model, while technology often replaces some jobs, it simultaneously generates new roles for humans to counterbalance the displacement effects of automation (Fleming et al., 2019).
Indifference and Stealing Like An Artist:
Standing on the shoulders of giants: I always draw inspiration from the concept of "stealing like an artist" in my projects. I strongly believe in the idea of "standing on the shoulders of giants to look further." I analyse the work of critically acclaimed designers not to steal their designs, but to understand their thought processes and the reasoning behind their decisions. I try to understand why some design decisions are good and others are not so good.
Indifference (Madvillainy):
Another important influence in this project is the idea of "Indifference," which is evident in the case of the album "Madvillainy" by MF DOOM and Madlib. In interviews, both artists have stated that they hate hip-hop and the culture around it. However, together they ended up creating one of the best hip-hop albums, "Madvillainy." This is attributed to their indifference to hip-hop. As they were not limited by the confines of hip-hop, they were able to take risks, innovate, and create a unique record. To surpass the giants, I wanted to look beyond the confines of graphic design.
3D Art:
This idea of indifference led me to explore the 3D art industry, which places a strong emphasis on creating systems that help artists produce art quickly and efficiently. This is mainly achieved through procedural workflows, unlike traditional digital design tools where a designer creates an asset layer by layer, pixel by pixel. In 3D art, procedural systems are developed so that artists can input various parameters. Based on these parameters, the system generates multiple results, allowing the artist to focus on the overall concept and aesthetic without having to worry about every single pixel of a design.
Practical Application
Incorporating semi-procedural methods into the graphic design pipeline:
So to keep up with the changes, the goal of this project was to incorporate semi-procedural workflows into the graphic design pipeline to take advantage of the benefits of the 3D art pipeline.
Case Study — Found Shapes:
To achieve the goal of incorporating semi-procedural methods into a graphic design pipeline, I've taken a previous project of mine called "Found Shapes," in which I transform typically disregarded elements such as cracks, stains, and puddles into visually captivating designs, breathing new life into these often overlooked features in our surroundings. I chose this project because its results are subjective and require human judgment to determine whether a composition is successful.
Tech Stack:
Blender: I used the software "Blender" to develop semi-procedural systems. Blender is an open-source software with powerful tools such as shader nodes and geometry nodes. These tools allow users to create semi-procedural systems that are limited only by their imagination.
Algorithms: I've created two algorithms using geometry and shader nodes:
The first algorithm scatters vectorised shapes based on procedurally generated noise patterns.
The second algorithm works like a "colour-by-number" activity, assigning random numbers to the scattered shapes and then assigning colours corresponding to the numbers based on a predefined palette.
Putting it to the test
Experiment brief: The developed tools were tested using a simple design brief. The goal was to create a visual that complements pre-set typography on a promotional banner with a pre-defined colour palette. This brief was designed to minimise design decisions, allowing me to focus solely on creating visually appealing compositions.
Traditional Method: Using traditional digital design methods, I was able to create 6 visuals in 120 minutes. On average, 10 minutes were spent on scattering shapes, and the remaining 10 minutes were devoted to colouring the composition.
Semi-procedural Method: Using the semi-procedural approach, I was able to create 6 visuals in 30 minutes. With the semi-procedural approach, I generated 750 unique visuals in 22.5 seconds, and then I spent the remaining 29 minutes and 37 seconds making creative decisions to select 6 out of the 750 variations.
Conclusion
Practical Results:
The semi-procedural process is 4 times faster than the layer-based process.
The semi-procedural process is 500 times more productive than the layer-based process.
Process Observations:
The major difference was observed in the overall design experience. When using the traditional method, I was making design decisions at every step. For example, when moving a shape, I was making a decision. This helped me create sound compositions, but it also caused tunnel vision, limiting me to explore only one idea at a time. On the other hand, when using the semi-procedural process, I was able to outsource the process of arranging and colouring and my role as a designer shifted from doing repetitive, laborious tasks to making creative decisions, almost like a creative director in the traditional graphic design pipeline and the algorithm was the junior designer exploring the different directions while I focused on the overall concept and aesthetic.
Other benefits:
The semi-procedural approach used in the experiment had an additional benefit. It allowed the developed system to be used beyond the scope of the "Found Shapes" project. This means that the system can be utilised to scatter any elements, enabling the creation of unique collages. As a result, this approach assists designers in creating tools that can be applied in various contexts, something not achievable through the layer-based approach. This ultimately enhances the value of the designer in the job landscape.
Final Thoughts:
After comparing the outputs from both approaches, the level of quality is virtually the same. Therefore, the semi-procedural approach not only offers practical benefits such as increased time efficiency but also leads to increased output. Additionally, this approach addresses the original goal of this project of shifting a designer’s focus from laborious tasks that can be automated by AI and machine learning, towards developing creative skills and insights.
Published
13 November, 2024
by
Param Pratap Walia
Future-Proof Design: Adapting Creativity for an AI-Driven World
This project explores how designers can adapt to the rise of generative AI by incorporating semi-procedural workflows into graphic design. Inspired by the efficiency of 3D art pipelines, the project tests this approach using Blender to create algorithms that automate repetitive tasks, freeing designers to focus on creative decision-making. A case study on "Found Shapes" demonstrated that the semi-procedural method is 4 times faster and 500 times more productive than traditional design processes. This shift redefines the designer's role, emphasizing creativity over labor-intensive tasks and enhancing versatility in various design contexts.
Introduction
This personal project is a response to concerns about how the recent rise of generative AI services could affect graphic design jobs. This is a scary idea as I am preparing to enter the workforce and I want to explore how I can adapt to this change.
Historical context of technology’s influence on the job landscape:
Technological breakthroughs have reshaped the job market by both eliminating and creating job roles. These innovations shift how work is divided between the tools we use (known as capital) and human workers. According to Acemoglu and Restrepo’s model, while technology often replaces some jobs, it simultaneously generates new roles for humans to counterbalance the displacement effects of automation (Fleming et al., 2019).
Indifference and Stealing Like An Artist:
Standing on the shoulders of giants: I always draw inspiration from the concept of "stealing like an artist" in my projects. I strongly believe in the idea of "standing on the shoulders of giants to look further." I analyse the work of critically acclaimed designers not to steal their designs, but to understand their thought processes and the reasoning behind their decisions. I try to understand why some design decisions are good and others are not so good.
Indifference (Madvillainy):
Another important influence in this project is the idea of "Indifference," which is evident in the case of the album "Madvillainy" by MF DOOM and Madlib. In interviews, both artists have stated that they hate hip-hop and the culture around it. However, together they ended up creating one of the best hip-hop albums, "Madvillainy." This is attributed to their indifference to hip-hop. As they were not limited by the confines of hip-hop, they were able to take risks, innovate, and create a unique record. To surpass the giants, I wanted to look beyond the confines of graphic design.
3D Art:
This idea of indifference led me to explore the 3D art industry, which places a strong emphasis on creating systems that help artists produce art quickly and efficiently. This is mainly achieved through procedural workflows, unlike traditional digital design tools where a designer creates an asset layer by layer, pixel by pixel. In 3D art, procedural systems are developed so that artists can input various parameters. Based on these parameters, the system generates multiple results, allowing the artist to focus on the overall concept and aesthetic without having to worry about every single pixel of a design.
Practical Application
Incorporating semi-procedural methods into the graphic design pipeline:
So to keep up with the changes, the goal of this project was to incorporate semi-procedural workflows into the graphic design pipeline to take advantage of the benefits of the 3D art pipeline.
Case Study — Found Shapes:
To achieve the goal of incorporating semi-procedural methods into a graphic design pipeline, I've taken a previous project of mine called "Found Shapes," in which I transform typically disregarded elements such as cracks, stains, and puddles into visually captivating designs, breathing new life into these often overlooked features in our surroundings. I chose this project because its results are subjective and require human judgment to determine whether a composition is successful.
Tech Stack:
Blender: I used the software "Blender" to develop semi-procedural systems. Blender is an open-source software with powerful tools such as shader nodes and geometry nodes. These tools allow users to create semi-procedural systems that are limited only by their imagination.
Algorithms: I've created two algorithms using geometry and shader nodes:
The first algorithm scatters vectorised shapes based on procedurally generated noise patterns.
The second algorithm works like a "colour-by-number" activity, assigning random numbers to the scattered shapes and then assigning colours corresponding to the numbers based on a predefined palette.
Putting it to the test
Experiment brief: The developed tools were tested using a simple design brief. The goal was to create a visual that complements pre-set typography on a promotional banner with a pre-defined colour palette. This brief was designed to minimise design decisions, allowing me to focus solely on creating visually appealing compositions.
Traditional Method: Using traditional digital design methods, I was able to create 6 visuals in 120 minutes. On average, 10 minutes were spent on scattering shapes, and the remaining 10 minutes were devoted to colouring the composition.
Semi-procedural Method: Using the semi-procedural approach, I was able to create 6 visuals in 30 minutes. With the semi-procedural approach, I generated 750 unique visuals in 22.5 seconds, and then I spent the remaining 29 minutes and 37 seconds making creative decisions to select 6 out of the 750 variations.
Conclusion
Practical Results:
The semi-procedural process is 4 times faster than the layer-based process.
The semi-procedural process is 500 times more productive than the layer-based process.
Process Observations:
The major difference was observed in the overall design experience. When using the traditional method, I was making design decisions at every step. For example, when moving a shape, I was making a decision. This helped me create sound compositions, but it also caused tunnel vision, limiting me to explore only one idea at a time. On the other hand, when using the semi-procedural process, I was able to outsource the process of arranging and colouring and my role as a designer shifted from doing repetitive, laborious tasks to making creative decisions, almost like a creative director in the traditional graphic design pipeline and the algorithm was the junior designer exploring the different directions while I focused on the overall concept and aesthetic.
Other benefits:
The semi-procedural approach used in the experiment had an additional benefit. It allowed the developed system to be used beyond the scope of the "Found Shapes" project. This means that the system can be utilised to scatter any elements, enabling the creation of unique collages. As a result, this approach assists designers in creating tools that can be applied in various contexts, something not achievable through the layer-based approach. This ultimately enhances the value of the designer in the job landscape.
Final Thoughts:
After comparing the outputs from both approaches, the level of quality is virtually the same. Therefore, the semi-procedural approach not only offers practical benefits such as increased time efficiency but also leads to increased output. Additionally, this approach addresses the original goal of this project of shifting a designer’s focus from laborious tasks that can be automated by AI and machine learning, towards developing creative skills and insights.
