From prompt to presence

Project information

Artificial intelligence is becoming increasingly present in creative design workflows. In transportation design, it can quickly generate large amounts of visual material and open unexpected directions for form exploration. This creates new possibilities, but also raises an important question: if images can be produced quickly and endlessly, where does the designer’s role begin?

This project investigates that question through a designer-led process. Instead of asking AI to create a finished car, the project uses AI-generated material as raw input for further development. The focus is on how the designer can remain in control by setting criteria, selecting what has potential, rejecting weak proposals, and translating useful visual qualities into a more coherent exterior form.

1. AI-assisted exploration

The process began with prompting and AI image generation. A large number of outputs were created to explore different proportions, silhouettes, surfaces, lighting conditions, and emotional expressions. This phase showed the strength of AI as an early exploration tool. It allowed the project to move quickly across many visual possibilities and discover directions that may not have appeared through traditional methods alone.

However, the amount of material also became a challenge. More images did not automatically create better design. In many cases, the outputs were visually attractive but too vague, too generic, or too unstable to develop further. This shifted the project from simply generating form to learning how to define, filter, and control it.

2. Filtering and authorship

One of the clearest learnings from the project was that authorship in an AI-assisted workflow is not maintained through prompting alone. It emerges through a series of decisions made after generation. The designer’s role became visible through choosing which outputs deserved development, rejecting visually seductive but weak proposals, and identifying which qualities supported the intended Peugeot-inspired direction.

Filtering therefore became an active design act. The project showed that selection is not a passive step, but a critical part of creating direction. AI could produce possibilities, but it could not decide which ones carried the right proportion, attitude, surface logic, or brand relevance. That responsibility remained with the designer.

3. Digital refinement and translation

Selected outputs were developed further through Photoshop, Vizcom, and Blender. These tools helped move the work from loose image inspiration toward a more controlled exterior design proposal. Proportions were adjusted, views were clarified, surfaces were refined, and visual ideas were translated into a more stable form language.

This stage also revealed a gap between image quality and design quality. A render could look convincing from one angle while failing in another. What seemed resolved in a curated digital image often needed to be reconsidered when translated across multiple views or into three-dimensional form. The process therefore became less linear than expected, moving back and forth between exploration, refinement, correction, and validation.

4. Physical validation through clay

Clay modelling became one of the most important parts of the project. It did not simply confirm the digital proposal; it challenged it. The physical model revealed where the front became too heavy, where the shoulders were too dominant, where the rear felt too weak, and where digital surfaces did not hold up in real space.

This changed the relationship between digital output and physical form. AI-generated images and digital renders could hide weaknesses through lighting, perspective, and atmosphere. Clay made those weaknesses visible. It helped test proportion, volume balance, surface tension, and overall presence in a way that the screen alone could not.

5. Learning from the workflow

The project showed that AI does not remove the designer’s work. It changes where the work happens. The challenge was not only to generate images, but to build a process strong enough to avoid drift, false novelty, and visually appealing but weak design proposals.

The designer’s value was found in defining criteria, interpreting ambiguous outputs, managing a non-linear workflow, protecting coherence across tools, and making decisions with intention. In this process, AI became useful when it was guided critically and used as support, not when it was treated as a replacement for design thinking.

6. Ethical and professional reflection

The project also raised questions about responsibility in AI-assisted design. These tools offer speed and range, but they can also encourage passive image consumption and shallow decision-making if the designer hides behind the output. Treating AI as a replacement rather than a support risks weakening the development of judgement, craft, and early-stage design thinking.

For this reason, the project argues for a selective and transparent use of AI. The role of the future transportation designer is not only to master new software, but to define opportunities, construct criteria, interpret outputs, and remain responsible for the final direction.

Outcome

The final outcome is a Peugeot-inspired exterior design proposal developed through a hybrid workflow between AI generation, digital refinement, and clay modelling. Due to the limited timeframe of the degree project, the proposal focuses on exterior form development, process exploration, and physical validation rather than full technical feasibility or production resolution.

Ultimately, From Prompt to Presence proposes a way of working where AI generates possibilities, the designer creates direction, and physical modelling tests the form in reality. The project highlights that the future value of the transportation designer lies not only in using emerging tools, but in knowing how to question them, guide them, and turn digital uncertainty into intentional physical form.