Google DeepMind has collaborated with renowned designer Ross Lovegrove and his studio to develop a fine-tuned AI model that can generate designs in Lovegrove’s distinctive style. The partnership also involved Creative Director Ila Colombo from Lovegrove Studio and design office Modem. The project used Gemini and Google DeepMind’s generative image technology to create a prototyping tool for the creative process.
The team chose to design a chair as their target project, citing it as a classic design challenge that balances fixed function with evolving form. To capture Lovegrove’s unique aesthetic, researchers fine-tuned their text-to-image model, Imagen, using a curated dataset of the designer’s personal sketches. The training focused on incorporating Lovegrove’s signature organic, fluid-like structures and biomorphic forms.
The development process emphasized collaboration between human designers and AI systems. Lovegrove Studio worked to decode and articulate the designer’s visual vocabulary, creating specific prompts to guide the model’s output. The team deliberately avoided using the word “chair” when prompting the AI, instead using creative synonyms to generate more diverse design concepts.
The collaborative effort ultimately produced a chair design that Lovegrove Studio considered an authentic extension of their work. The final concept was manufactured as a physical prototype using metal 3D printing technology. “For me, the final result transcends the whole debate on design. It shows us that AI can bring something unique and extraordinary to the process,” said Ross Lovegrove.
Source: blog.google