Brief of UCL AI + Design Workshop

date

type

status

slug

summary

AI + Design Workshop of RC16: Towards Non-anthropocentric Intelligence in Design

In the era of exponential technological advancement, contemporary urban challenges demand paradigm shifts in both thinking and methodological approaches. This workshop proposes a pioneering framework that transcends traditional anthropocentric design paradigms by synthesizing multiple forms of intelligence - artificial, biological, and human - into a cohesive hybrid intelligence system for design innovation.

We position AI not merely as a tool, but as a collaborative agent in the creative process, facilitating a non-anthropocentric design methodology that expands our problem-solving capabilities beyond inherent human cognitive boundaries. This workshop introduces advanced AI platforms across the spectrum - from open-source frameworks like ComfyUI to commercial solutions such as Midjourney, chatGPT, Tripo3D, runway - to generate sophisticated multimodal outputs including text, imagery, video and three-dimensional prototype.

The course emphasizes the development of nuanced prompting strategies and computational workflows that enable meaningful dialogue between human designers and AI agents. Students will learn to orchestrate complex systems in ComfyUI for precise control over generative processes, while developing an understanding of how biological intelligence can serve as both an inspiration source and a mediating protocol between artificial and natural systems. This tripartite integration of human insight, AI capabilities, and biological principles opens new pathways for addressing complex urban and environmental challenges.

Course Outline

Course 1 - Lecture : AI + Design introduction & Text to image basic

AI Toolkit collection introduction.
Install and utilization of ComfyUI.
Skill of Prompt engineering.
Customized (lora) model and extension installation.

Course 2 - Lecture : AI based image to image generation.

Basic of image to image translation with example of biological experiment.
Image translation with Controlnet / Ipadapter.
Image AI post-process: super resolution, in-painting, out-painting.
Guideline of draft submission.

Course 3 - Tutorial: Review of AI + Design practice

Summit PDF draft about integrating AI with your present research project.

Course 4 - Lecture : Deep dive of ComfyUI

Apply LLM / LVM (Large language / Vision model) of COT (chain of thought) / AI agent in ComfyUI.
Connect LLM / LVM with Image generation model.

Course 5 - Lecture: Video and 3D generation with AI

AI Video generation with online platform: runway \ sora \ klingai.
AI 3D generation with with online platform: meshy. Ai \ tripoAI \ hyper3d.ai \ CSM \ zoeDepth.
Finetune the texture of the existing 3D model.

Course 6 - Tutorial: Mid-term Review of AI + Design practice.

Course 7 - Tutorial: Final review before workshop submission.

Expected Learning Outcomes

Students will develop proficiency in hybrid intelligence design methodologies, demonstrating the ability to orchestrate meaningful collaborations between human creativity, AI systems, and biological principles in their design process.

Students will master advanced workflows in ComfyUI and other AI platforms, enabling them to generate and refine multimodal outputs that respond sensitively to site context and environmental parameters.

All workshop outputs (text, imagery, video, 3D models, parameters, etc.) will be curated into a comprehensive portfolio that demonstrates both technical competency and theoretical understanding of non-anthropocentric design approaches.

Students will develop critical perspectives on the role of AI as collaborative agents in design, moving beyond tool-based applications to establish new paradigms for hybrid intelligence creation.