Google DeepMind has collaborated with famend designer Ross Lovegrove and his studio to develop a fine-tuned AI mannequin that may generate designs in Lovegrove’s distinctive model. The partnership additionally concerned Artistic Director Ila Colombo from Lovegrove Studio and design workplace Modem. The mission used Gemini and Google DeepMind’s generative picture expertise to create a prototyping instrument for the inventive course of.
The crew selected to design a chair as their goal mission, citing it as a traditional design problem that balances mounted operate with evolving kind. To seize Lovegrove’s distinctive aesthetic, researchers fine-tuned their text-to-image mannequin, Imagen, utilizing a curated dataset of the designer’s private sketches. The coaching targeted on incorporating Lovegrove’s signature natural, fluid-like buildings and biomorphic varieties.
The event course of emphasised collaboration between human designers and AI programs. Lovegrove Studio labored to decode and articulate the designer’s visible vocabulary, creating particular prompts to information the mannequin’s output. The crew intentionally prevented utilizing the phrase “chair” when prompting the AI, as an alternative utilizing inventive synonyms to generate extra various design ideas.
The collaborative effort finally produced a chair design that Lovegrove Studio thought-about an genuine extension of their work. The ultimate idea was manufactured as a bodily prototype utilizing metallic 3D printing expertise. “For me, the ultimate outcome transcends the entire debate on design. It reveals us that AI can carry one thing distinctive and extraordinary to the method,” stated Ross Lovegrove.
Supply: weblog.google