Manas Bhatia
Table of Contents
1 – Why?
From Conflict to Consensus
AI-Driven Engagement for Community-Centered Urban Design.
Before we dive into this project, I want you to imagine a site/area in the city near you that you really like. A place that feels like it belongs to you. How would you feel if a developer constructed an out-of-place building there? Would you feel voiceless in that situation?
Let's consider another case. Consider our studio space. The long table is often occupied by other departments, leaving little room for CDP students. While not all of us may need the table, each of us likely has strong opinions on how that space should be used.
These examples illustrate spatial conflict reflecting how we all have unique visions for the spaces we inhabit.
In NYC, this conflict is seen on a larger scale. So I began by exploring Community Profiles and district needs statements. A recurring theme emerged: challenges around affordable housing and parks and open spaces are common across many neighborhoods.
One specific example is the Elizabeth Street Garden, where local users want to preserve the garden as a green space, while the city plans to replace it with affordable senior housing. This tension between preserving public spaces and meeting housing demands has led to protests and legal battles. Such conflicts show a disconnect between developers, city planners, and local communities. As cities rapidly grow, the voices of everyday residents often get drowned out.
2 – What?
Research Question
Current observations
In rapidly urbanizing cities, the voices of local communities frequently get lost in the clamor of large-scale development, leading to spaces that overlook everyday needs and desires.
At every scale, from individual streets to entire neighborhoods, these misalignments fuel tension and disconnect between stakeholders.
The disconnect between architects, urban planners, and local communities as well as diverse urban experiences has highlighted the need for direct engagement and collaboration with city residents.
"People see the cities in different ways"
Key Questions
How can we create cities that truly reflect the wishes of their people, especially when conflicting interests—government, developers, and local residents—often shape the urban landscape?
How might we use AI to bridge that gap, empowering residents to actively participate in the design of their environment?
Proposed Hypothesis
Having a digital platform/tool that allows local users voice their opinion by visualize their proposals/suggestions using Gen AI would improve overall community participation in the urban design process.
"Through this project, I want to build a world where cities are shaped not just by experts, but by the people who live in them every day."
3 – How?
City.AI
Visual AI tool for civic engagement
This project envisions a future where AI bridges this divide by enabling communities to contribute meaningfully to urban planning. City.AI is a web-based platform with two main tools:
- "INSIGHT": An interactive map for sharing aspirations and understanding local needs.
- "UNIFI": Uses AI to generate and visualize 3D proposals from community inputs.
By capturing local perspectives and creating a collaborative design process, the project challenges traditional top-down approaches to urbanism. It explores multiple scales from streets to neighborhoods and enables practices that align design decisions with collective needs. The tool can be used to solve conflicts between multiple stakeholders and scaled for site, to neighbourhoods, to cities.
3 – Methods
Computational Tools and Methods
This project leverages AI-driven tools to facilitate community participation in urban design. The computational methods include text-to-3D modeling, interactive mapping, and AI-powered data analysis. These tools allow users to input aspirations, generate 3D visualizations of their ideas, and see their impact on the city in real-time. By using generative AI for modeling and machine learning to analyze trends in community inputs, the project creates an iterative feedback loop. This workflow democratizes urban planning, enabling non-experts to contribute effectively. The methods also address the scale and complexity of urban environments, offering scalable and adaptable solutions for diverse urban contexts.
Design Methods
"Participatory, Speculative, and Experimental."
By involving residents directly in the design process, it shifts the narrative from top-down planning to a collaborative framework. The speculative approach enables communities to imagine alternate futures, while the participatory aspect ensures their voices shape these visions. Experimental workflows with AI and generative tools make it possible to visualize and prototype ideas quickly. This methodology aligns with activist practices, critiquing the exclusion of local voices in traditional planning and proposing a model where everyone is empowered to co-create urban spaces responsive to their needs.
5 – Precedents
Inspiration
UrbanistAI
Urbanist AI provides a platform for residents to visualize and propose urban design solutions using generative AI. This aligns closely with Unifi, as both utilize AI to empower users in creating 3D visualizations of their design ideas, democratizing access to urban planning tools.
Commonplace
Commonplace facilitates community engagement by collecting resident feedback on urban developments through interactive maps and visual tools. Its emphasis on connecting planners and residents highlights the collaborative potential of Insight in gathering and reflecting local perspectives.
KPF HHL
The Hawaii Housing Lab (HHL) by KPF uses a web-based tool to visualize housing data and analyze community input to inform better planning decisions. Its focus on integrating diverse data sets and feedback inspires Insight’s interactive map, which captures and visualizes community aspirations at scale.
Maptionnaire
Maptionnaire enables participatory mapping, allowing communities to share insights and feedback about their surroundings. This approach informs Insight’s ability to collect hyper-local data, ensuring that design decisions are rooted in the specific needs and aspirations of the people they impact.
5 – Audience
Who does it impact?
The primary audience includes urban residents, community groups, urban planners, and local governments. The platform is designed to empower underrepresented voices in urban planning, particularly those excluded from traditional processes. For alternative audiences, such as children or non-digital natives, simpler interfaces or in-person workshops could make the platform more inclusive. Expanding accessibility through multilingual support and offline modes could further broaden the audience.
5 – Data
Data
The project relies on community input text prompts/inputs describing what they want to visualise as well as the final design or image that they submit. Early analyses of user-provided aspirations and site-specific features demonstrate how data informs design decisions. For testing the workflow, XYZ site in the Columbia campus was selected, the users were students from GSAPP, and a prototype dataset of the visuals and the prompts, was created and mapped, showing how AI-driven insights can support collaborative planning.
5 – Materials
Materials/Sensors/Algorithms
The project uses open source generative AI Image generation models from Huggingface, LLMs (Large Language Models) specifically GPT by OpenAI.
In the future we can also integrate AI 3D generation instead of relying on just images. That way the user can also see the visualization from different angles to get a better understanding. We can also integrate VR into this tool, as well as real time rendering AI models, so that users can also experience their designs spatially and virtually.
5 – Bibliography
Some articles to read
Hsu, Y.-C., Huang, T.-H. K., Verma, H., Mauri, A., Nourbakhsh, I., & Bozzon, A. (2022). Empowering local communities using artificial intelligence. Patterns, 3(3), 100449. https://doi.org/10.1016/j.patter.2022.100449
Williams, Sarah, Sara Beery, Christopher Conley, Michael Lawrence Evans, Santiago Garces, Eric Gordon, Nigel Jacob, and Eden Medina. 2024. “People-Powered Gen AI: Collaborating with Generative AI for Civic Engagement.” An MIT Exploration of Generative AI, September. https://doi.org/10.21428/e4baedd9.f78710e6.
Bibri, S. E., Huang, J., Jagatheesaperumal, S. K., & Krogstie, J. (2024). The synergistic interplay of artificial intelligence and digital twin in environmentally planning sustainable smart cities: A comprehensive systematic review. Environmental Science and Ecotechnology, 20, 100433. https://doi.org/10.21428/e4baedd9.f78710e6.
Bloomberg News. (2024, March 13). AI-powered urban innovations bring promise, risk to future cities. Bloomberg. Retrieved from https://www.bloomberg.com/news/articles/2024-03-13/ai-powered-urban-innovations-bring-promise-risk-to-future-cities
The project People-Powered Gen AI: Collaborating with Generative AI for Civic Engagement by Sarah Williams explores how AI can be leveraged to analyze qualitative data for civic purposes, while the Bloomberg article AI-Powered Urban Innovations Bring Promise, Risk to Future Cities highlights the potential and challenges of integrating AI in urban innovations. Similarly, the review by Simon Elias Bibri and colleagues discusses the interplay of AI and digital twins in planning sustainable smart cities, and Yen-Chia Hsu et al.’s work emphasizes empowering local communities through AI. Together, these works inspire my project by showcasing how AI can be used to analyze complex datasets, engage communities, and enhance sustainability efforts in urban environments.