In seiner Funktionalität auf die Lehre in gestalterischen Studiengängen zugeschnitten... Schnittstelle für die moderne Lehre
In seiner Funktionalität auf die Lehre in gestalterischen Studiengängen zugeschnitten... Schnittstelle für die moderne Lehre
The Bauhausifier is a project developed as part of the 'Almost There: Bauhaus Jubilee' elective, guided by Prof. Dr. Uwe Gellert.
For the Bauhaus 100 Years Jubilee, we developed an AI-powered interactive installation to transform human movement into real-time Bauhaus-inspired architectural models. Utilizing and training a custom AI model on Bauhaus data, the system processed live camera feeds to generate structures locally, without requiring an internet connection. Another feature integrated ChatGPT’s API with ElevenLabs’ voice model, providing participants with a playful critique of their poses. Showcased at the Dessau Design Show, the installation attracted visitors from around the city to engage in an interactive experience, blending AI, design, and Bauhaus principles in an innovative way.
The project began with an in-depth study of the Bauhaus manifesto and history, including research at the Bauhaus library, where we explored the works of Bauhaus masters to understand their design philosophies. This research helped us envision a futuristic reinterpretation of Bauhaus principles. After brainstorming sessions, we decided to integrate AI to create an interactive installation that brings these ideas to life through modern technology.
We conducted extensive research on embedding Bauhaus concepts into an engaging, AI-driven experience. The technological development involved testing various AI models and tackling challenges such as integrating the camera feed with AI processing and refining the model to generate meaningful Bauhaus-inspired forms rather than hallucinations. Through iterative improvements, we successfully developed a locally trained AI prototype, specifically designed to bring this vision to life.
Building the prototype technology stack involved integrating a camera, a large display, and a locally trained AI model running on my own PC to create a seamless, immersive experience.
The core of the system was a Stable Diffusion model, fine-tuned on Bauhaus data and running locally to ensure real-time processing, and full privacy without internet dependency. The camera captured users' poses, which were then analyzed and transformed into Bauhaus-inspired architectural models displayed on a big screen. To enhance immersion, we designed a partitioned space within the room, allowing users to step into a dedicated interaction zone where they could fully engage with the AI-driven experience.
In the end we achieved the goal, creating an interactive installation that leverages the latest AI technology to provoke thoughts about Bauhaus principles, reinterpreting its aesthetics and philosophy through modern computational tools and engaging the visitors to become part of a live artwork creation process.
The project was showcased at the Dessau Design Show, where it received an enthusiastic response. Visitors enjoyed experimenting with different poses to generate unexpected Bauhaus-inspired results, from abstract artworks to architectural forms, making it a highly engaging and interactive experience.