The world is undergoing a transformative shift because of artificial intelligence (AI) that's affecting everything in our lives and in different industries. This is nothing sort of generational transformation like when the internet was created or when the light bulb was invented. Among the domains experiencing a profound impact, the automotive industry stands prominently, with design emerging as a pivotal focal point.
At first, people were doubtful whether AI could be creative and handle tasks in fields like design, music, and art. Many thought these areas would not be affected by this technology. But now, in the year 2023, things are changing. These AI tools are being used/adopted in vehicle design and other creative tasks. This is just the beginning of a big transformation, where AI is set to revolutionize how vehicles are designed and will become an essential tool in the process.
Before we talk about how these technological advancements will be used in modern automotive design, we need to emphasise how design practices have changed over the years.
Automotive Design Evolution
Automotive design has changed a lot over the years, with each decade bringing new transformations. In the 1990s, groundbreaking technologies like Photoshop and Computer-Aided Design (CAD) became important. Visionary designers understood the value of these tools, and over time, they became essential parts of the design process. In the dynamic world of car design, where innovation is crucial, these technological advances have played a key role in shaping the creative landscape.
In the 1990s, the introduction of Photoshop changed how designers worked. Designers could now quickly try out multiple design ideas for one project within a few hours, a big difference from the constraints of traditional methods. Even though hand-drawn sketches kept their artistic value, Photoshop became an additional tool that provided remarkable control over visual communication and contributed to improving design concepts.
Similarly, the introduction of Computer-Aided Design (CAD) marked a new era in automotive design. Designers could now create detailed 3D models in a few weeks, allowing them to see designs in a realistic setting even before committing to a specific concept. This speeds up decision-making, helping design teams make informed choices efficiently. It's important to note that CAD didn't replace traditional methods like clay modelling in many studios but instead smoothly connected the gap between initial sketches and physical prototypes. While speaking to Pontus Fontaeus (Executive Design Director, GAC) about his journey into CAD he mentioned “I supported the use of digital tools in design studios, moving away from traditional 3D assessments and clay modelling as in the past, especially during tough economic times, we had to find more efficient ways. This started with Volvo, where, due to financial constraints, we had to discover innovative shortcuts”.
The use of technology has continued to evolve, including recent advancements like Virtual Reality (VR) and, importantly, Artificial Intelligence (AI). While some doubt the idea of a computer and algorithms independently designing a car, the real role of AI is not in actual design but in generating creative ideas. It's worth exploring how AI is currently being used in the design process.
Role of AI in Automotive Design
Currently, AI is used to create design ideas at the beginning of a project during the research and ideation phase. AI acts as a tool to generate ideas, simulating the approach of a young designer. These ideas are more on the creative side rather than polished or fully developed. It's up to the designer and the team to choose ideas and develop them according to their preferences. This is where the team focuses on new creative ideas, and so far, AI has been quite effective at generating ideas that are unconventional but interesting.
This leads us to discuss the fundamental reason for using AI. Automotive designers have consistently sought new inspirations to create fresh forms, shapes, and even emotions in their designs. They draw inspiration from various sources like architecture, yachts, planes, animals, and birds. Looking beyond vehicles to universal objects has been crucial in generating interesting ideas and driving innovation in automotive design. Discovering something new is a crucial aspect of design and determines how we perceive a design as modern, recent, or old.
This is where AI comes into play, serving as an excellent tool for generating hundreds of ideas quickly. It introduces a new creative dimension to automotive design, offering significant potential for creating ideas without spending extensive time on sketching, rendering, and researching. This marks a significant advancement in pushing automotive design into a new era. The most interesting part of this new technology is that the generated ideas can be intriguing, fascinating, and often outside what a designer would typically think of.
To understand more we asked Pontus about his thoughts on integrating Artificial Intelligence (AI) in the design process at GAC. He explained, “In our studio, we start with traditional methods and later use AI to explore different options. It's important not to dive for quick solutions because discomfort and challenges in the design process often result in the best outcomes.
"AI should play a supporting role in our creative efforts, not the lead" - Pontus Fontaeus
Full interview with Pontus Fontaeus - Executive Design Director, GAC USA
How does AI create images?
After discovering the tool's effectiveness in generating images, the first question that comes to mind is How. It is the most interesting part of the process, and the way designers can make use of this technology can entirely alter the outcome while offering a lot of control. To gain more insights, we asked Jo Stenuit (Design Director, Mazda Motor Europe) about the potential threat of AI in generating content independently, and he clarified that AI doesn't generate random results. Designers can improve and control the outcomes by using the significant control offered by certain AI tools. Ultimately, the quality of the design output depends on the quality of the input. The resulting work reflects the skills of the designer in providing excellent information or clear guidelines for AI-assisted iterations. This situation is similar to the early days of Photoshop when some individuals downloaded the software for free, mistakenly thinking they could use it effectively.
Full Interview with Jo Stenuit - Design Director, Mazda Motor Europe
Currently, there are two ways to use AI for design: one with text input and the other with sketch input. Tools like Vizcom can handle both sketch and text input. This allows users to control the initial image, influencing the basic shape of the car. Adding text input can affect various aspects like rendering style, colour choice, environment, and design parameters such as body style or design language. There are also other tools like Midjourney and Stable Diffusion, which use text input to create images. Using AI and generating images through proper prompts is a skill. Many believe that having good research skills, including using the right prompts, will be crucial in the future.
Vizcom Interface - Sketch and Text based input
Mid Journey Interface - Text based input
What's interesting is another factor: Custom AI Models
When Stable Diffusion is fully installed on a machine with a graphics card like the Nvidia RTX series, it can create AI images directly on the device without needing an internet connection or sending data to servers. The user also needs to install a set of images that the AI is trained on. Each image is tagged with certain attributes, like soft, sharp, or elegant. The AI then uses these images as references, referred to as a model.
This is where the next generation of AI tools will thrive. Manufacturers and design teams can build their own AI models by feeding thousands of sketches created by their designers over the years. This model can be customized to be entirely relevant to a specific brand. This makes AI more personalized and gives the user greater control. Designers can manage the tags, and images, and guide AI decisions in a particular direction, leaving the rest to the algorithm. These models have even enabled brands to optimize the shape for aerodynamics by inputting a side-view sketch of a car.
Conclusion
The integration of AI into automotive design represents a transformative addition to the designer's toolkit, ready to enhance rather than replace the human touch. This innovative tool promises to enhance the overall value of the design process, encouraging increased creativity and efficiency. Without replacing the designer, AI acts as a catalyst, lifting creative output and streamlining the workflow. It also brings a notable advantage by substantially reducing the time and human resources needed to generate numerous ideas. Designers using these tools can quickly create thousands of concepts, significantly speeding up the ideation phase. Consequently, individuals skilled in using AI tools are positioned to hold a distinct professional designation, acknowledging the specialized skills needed to fully utilize these cutting-edge technologies.
In simple terms, the collaboration between human designers and AI creates a synergistic dynamic, where each one's strengths contribute to a more robust and innovative design process. AI speeds up idea generation and adds a layer of complexity and creativity that meets modern design demands. As a result, the combination of human intuition and AI capabilities marks a shift in the automotive design landscape. This signifies a future where the fusion of human expertise and technological prowess leads to unprecedented advancements in creativity and efficiency.