“Housing Diversity Means Diverse Housing”: Blending Generative AI into Speculative Design in Rural Co-Housing Communities

Research Background and Issues

What problems or challenges did the authors identify?

1. Complexity of Environmental and Social Issues: Environmental sustainability has evolved into a "wicked problem" that requires collaborative exploration among multiple stakeholders to identify solutions.
2. Limited Understanding of Future Visions for Cohousing Communities: Existing research primarily focuses on observing current practices in cohousing communities, neglecting how community members plan and envision the future.
3. Limitations of Technology-Driven Sustainability Design: Traditional technology-driven solutions often overemphasize behavioral metrics while overlooking complexity and associated values.

Why is this issue important?

1. Addressing Global Challenges: Cohousing communities possess unique advantages in resource sharing, social cohesion, and ecological regeneration, offering innovative models for sustainable development.
2. Driving Social Change: Exploring future cohousing models not only helps address housing issues but also challenges traditional perceptions of housing models and resource distribution.

Research Motivation and Related Work

1. The authors aim to integrate Generative AI (GenAI) into speculative design to inspire diverse thinking about the future of cohousing communities. Speculative design focuses on raising questions, avoiding oversimplification of complex issues.
2. Drawing on existing research, the authors seek to go beyond traditional "persuasive technology" approaches in sustainability studies, encouraging open dialogue and contextual understanding.
3. The authors aim to address the following research questions:
   • RQ1: How can GenAI help uncover the values and motivations behind community members' behaviors?
   • RQ2: What are the key insights from these findings for supporting sustainable Human-Computer Interaction (HCI) design?

Solutions

What methods or solutions did the authors propose?

1. Designing Speculative Tools: Developed a tool integrating Generative AI to help community members envision future scenarios.

2. Research Design Phases: Structured the workshops into three phases:

   • Uncovering Challenges and Experiences
   • Generating AI-Produced Scenarios and Discussions
   • Backcasting and Provocation

3. Tool Implementation: Created an online platform combining large language models (e.g., OpenAI GPT-3.5) and diffusion models (e.g., MidJourney) for text descriptions and image generation.

What are the innovative aspects of this solution?

1. Enhanced by Generative AI:
   • Leveraged AI to generate textual and visual content for future community scenarios.
   • Enabled iterative modification of generated content, enhancing participant interactivity.
2. Human-Centered and Contextual Design:
   • Incorporated perspectives of specific community members (e.g., long-term residents, activists) into the design process.
   • Emphasized exploring sustainability from individual motivations and emotions.

What are the implementation steps?

1. Recruiting Participants: The study involved 14 participants, including long-term residents and advocates of cohousing communities.
2. Using the Tool: Participants generated future scenarios by inputting keywords and descriptions, refining them through discussions.
3. Exploring Backcasting: Participants reflected on how to achieve ideal scenarios from the present, raising critical questions.

Research Findings

What specific outcomes were achieved?

1. Identified five core themes:

   • Diverse Housing Forms: Challenged traditional fixed housing models, exploring mobile housing options (e.g., RVs, dome houses).
   • Conflict as an Opportunity for Growth: Viewed conflicts within communities as opportunities for learning and building trust.
   • Decentralized Small Entities: Promoted efficient resource use through small-scale production and shared facilities.
   • Social Acceptance: Gained recognition and support through interactions and activities with neighboring communities.
   • Reducing Structural Constraints, Focusing on Essentials: Prioritized optimizing resource allocation and design processes based on fundamental values.

2. Highlighted the potential of Generative AI in supporting community members to envision future scenarios, such as visualizing complex ideas and sparking discussions.

How does it compare to existing solutions?

1. Speculative design tools powered by AI can rapidly generate scenarios, broadening participants' perspectives.
2. Encouraged open discussions and questioning of existing rules, breaking the limitations of traditional solutions.
3. Emphasized the role of personal stories and emotions in sustainability design, rather than relying solely on macro-level models.

What were the experimental or evaluation results?

1. Most participants used GenAI to create future community scenarios that combined personal experiences with reflections on societal and cultural issues.
2. During the design workshops, participants deepened their understanding of sustainability and community design by decoding AI-generated texts and images.

Limitations and Future Directions

1. Limitations:

   • Participants were primarily residents of remote areas in Australia, with most over 50 years old, potentially limiting diversity of perspectives.
   • The study focused mainly on scenario design and did not extend to the development of real-world prototypes.
   • Initial generated scenarios sometimes lacked creative diversity.

2. Future Directions:

   • Conduct research across broader geographic regions and age groups.
   • Explore how to integrate LLMs with contextual information to improve the relevance of generated results.
   • Develop practical technological and tool prototypes based on community scenarios.

Through this research, the authors provide a promising framework for designing cohousing communities and demonstrate the significant role of Generative AI in facilitating scenario exploration and inspiring new ideas for sustainable development.