AI in Furniture Design: A New Blueprint from Lima

At Ricardo Palma University, a CNC router carved plywood components from neural networks. The inaugural Design + AI Hackathon, held from September 22–24, invited participants to prototype furniture informed by artificial intelligence. The challenge was to use AI as a collaborator in rethinking form and function.

Organized by the Faculty of Architecture and Urbanism, the hackathon paired students and professionals from design, engineering, and computer science. A cohort of 47 worked in five teams to integrate generative AI models with hands-on fabrication. One team trained a neural network on a dataset of mid-century modern chairs, developing a bench that combined curved wooden slats with algorithmically shaped seats. Another proposed a shelving unit balancing structural efficiency with aesthetic tension.

Ricardo Ramos, a professor of digital fabrication at the university, explained the event’s focus on furniture: “AI in architecture tends toward abstraction—parametric skylines or idealised facades. Furniture, in contrast, is intimate. These objects are touched, lived with. AI can propose forms that challenge our assumptions but must remain human-scaled.”

This duality of AI as speculative and pragmatic was evident in the hackathon’s constraints. Each team received a defined material palette: locally-sourced plywood, metal rods, and foam. While generative models dictated geometry, teams ensured their designs could be fabricated in the campus workshop. A digital-physical loop emerged, testing AI outputs against the constraints of saws and drills.

The jury included Maria Luisa de la Torre, a Peruvian furniture designer, and Javier Cuellar, who leads AI integration at a Lima-based tech startup. De la Torre highlighted the bench project as exemplary: “It doesn’t just use AI to generate ornament. The algorithm affected joinery, ergonomics. It’s both forward-looking and deeply grounded in craft.”

The event also revealed limitations. One team attempted to use ChatGPT to optimize plywood cutting diagrams but found its suggestions too generic. Another struggled with the technical challenges of importing AI-generated meshes into CAD software. Ramos views these missteps as part of the process: “AI is not a shortcut. It demands iteration, translation. That friction is creative.”

Although rooted in experimentation, the hackathon gestures toward commercial potential. Cuellar noted that Peruvian furniture makers often lack access to advanced fabrication tools. Events like this could bridge that gap by demonstrating how AI might increase efficiency or inspire new product lines. Yet he cautions against over-automation: “Our industry values craft. AI should augment, not replace that.”

The 1st Design + AI Hackathon positions Lima as an emerging node in global conversations about technology and design. Institutions like MIT Media Lab or London’s Bartlett School of Architecture often dominate this terrain. Ricardo Palma’s event is distinct in its regional focus. Materials were sourced locally; the prototypes reflect contexts where resources can be limited, and designs must be adaptable.

The prototypes will be exhibited at the university’s annual design showcase in November 2023. Whether any evolve into production pieces remains unclear. For Ramos and his students, that ambiguity is part of the attraction. “Design is not just solutions. It’s questions,” he said, standing next to a half-assembled frame whose curves he described as “accidental but intentional.”

In a field increasingly shaped by algorithms, events like these remind us that innovation is iterative, contested, and local.