The project started with an abstract exploration phase: AI was used to cross-combine diverse art influences and produce raw, non-literal form studies. As the research converged, the final reference system was focused on Olafur Eliasson, because his work translates perception into form—light gradients, shadow control, reflections, and spatial illusion. These qualities created the strongest link to Maybach and supported a distinct, contemporary form language.

Alongside the experimental research, Maybach heritage acted as a structural reference. Concours d’Elegance vehicles informed the sense of presence: long, calm proportions, precise detailing, and controlled reflections. This anchor ensured the concept stayed unmistakably Maybach—while still allowing a new, contemporary form language to emerge.

In parallel with installation-focused experiments, I used AI to generate abstract, art-influenced automotive compositions as a form-research tool. These outputs were treated as raw material—then filtered, edited, and translated into design-relevant cues through sketching and surfacing logic. Running multiple threads simultaneously helped map the design space, before the project ultimately converged on the Olafur Eliasson–driven direction as the final reference system. This section documents the broader exploration that informed the final language.

Although the diploma project concluded in 2024, I revisited it in 2025 as an independent continuation. Using the original digital 3D model as a foundation, I trained a Stable Diffusion model to regenerate the vehicle consistently and to cross-combine it with existing Mercedes vehicles. This served as a direct feedback loop: a practical test of how transferable and “Mercedes-compatible” the new form language is when pushed into real brand contexts.

This project confirmed that AI can expand the search space dramatically—especially in early form research—but it does not replace the craft of vehicle design. When outputs look too realistic too early, they can “lock” decisions visually and reduce the willingness to reshape fundamentals. To avoid generic outcomes, the key is to aim for non-automotive abstraction first, then translate it intentionally into automotive cues through designer-led interpretation. In practice, the strongest results came from the hybrid loop: AI for hypotheses, and traditional design skills for filtering, editing, and committing those hypotheses into a coherent, ownable form language.

Beyond speed, the real value of AI is contrast: it helps reveal what is distinctive and what is accidental by generating many alternatives quickly. But authorship still depends on decision-making—what to keep, what to remove, and what to push into a clear design statement. Keeping the process iterative—generate, select, reinterpret, and validate against proportion, brand logic, and surface integrity—prevents the work from becoming “AI-polished” but design-empty. Used this way, AI becomes a collaborator in exploration, while the designer remains responsible for meaning, coherence, and credibility.