When Your Home Knows You: AI and the Future of Personalised Living | Shuai Lu | TEDxTsinghua SIGS

Key Concepts

• Generative AI & LLMs: Tools used to translate natural language requirements into architectural designs.
• 3D Concrete Printing: A robotic construction method that eliminates the cost penalty of non-standardized building designs.
• Human-Centric Smart Environments: Using wearable physiological data to automate climate control.
• Adaptive Architecture: Buildings that physically change shape or configuration to optimize environmental comfort.

---

1. The Problem: The "Normal" Building Paradigm

Professor Shrayer Liu argues that modern urban living is defined by "normal" buildings—standardized, mass-produced units that fail to account for individual human diversity. Whether in high-density cities or suburban UK, the "copy-paste" approach to architecture forces diverse individuals into identical spaces, compromising personal preferences regarding layout, light, and functionality.

2. Personalized Design via Generative AI

Liu’s research group is developing tools that leverage Generative Models and Large Language Models (LLMs) to democratize architectural design.

• Methodology: Users input natural language requirements (e.g., "120 square meters," "60 square meter living room for parties").
• Functionality: The AI generates compliant designs that adhere to regulatory guidelines. Users can iteratively refine these designs through conversational prompts (e.g., "add a balcony to the north").
• Impact: This removes the initial barrier of needing a professional architect to conceptualize a custom home.

3. Affordable Customization: 3D Printing

A major challenge to personalized architecture is the high cost of non-standard construction. Liu proposes 3D Concrete Printing as the solution:

• Process: Robots mounted on rails move across a construction site, laying concrete layer by layer.
• Economic Advantage: Because the robot is programmed rather than manually laboring, the cost of building a unique house is nearly identical to building a standard one.
• Social Impact: Construction becomes faster (finishing in days) and safer, shifting the role of human workers from heavy, repetitive labor to supervisory or observational roles, effectively turning the construction site into a more "human" and "joyful" environment.

4. Intelligent, Responsive Environments

Beyond the physical structure, Liu addresses the "dumb building" problem—where homes do not adapt to the occupants' needs.

• Physiological Feedback Loops: By integrating wearable devices (smartwatches) that track heart rate, skin conductivity (EDA), and skin temperature, AI can infer a user's thermal comfort preferences.
• Application: The AI uses this data to automatically adjust HVAC systems, resolving conflicts between occupants with different temperature needs.

5. Adaptive Architecture

The final frontier discussed is the physical adaptability of the building itself. Liu showcases two prototypes:

• Umbrella Pavilion: A structure that changes shape based on environmental conditions to maintain comfort.
• Intelligent Climate Interface: A window system inspired by traditional Chinese design that adjusts its configuration to optimize light and airflow based on real-time environmental data.

6. Synthesis and Conclusion

Professor Liu concludes by emphasizing that while virtual reality and AI are powerful tools, the physical reality of our living spaces remains the primary determinant of our quality of life. The goal of integrating these technologies is to create homes that are:

1. Personalized: Reflecting individual needs rather than standard units.
2. Beautiful: Utilizing advanced design and construction techniques.
3. Environmentally Friendly: Through adaptive, responsive systems.

Notable Quote:

"I'm not crazy about reality, but it's the only place to get a decent meal... I also equally, if not more, love the reality. Love the [space] we live [in] because I think that's the thing [that] really determines our living quality." — Shrayer Liu