When design workshops meet chatbots: Meaningful participation at scale?

Abstract

This paper explores the potential of chatbots, powered by large language models, as a tool for fostering community participation in architectural and urban design. By taking a hybrid approach to community participation in a real-world mixed-use building project, in which we integrated remote chatbot engagements with face-to-face workshops, we explored the potential for a hybrid approach to scaling up the reach of participation while ensuring that such participation is meaningful, genuine, and empowering. Our findings suggest that a hybrid approach amplified the strengths and mitigated the shortcomings of the two methods. The chatbot was effective in sustaining the length of participation, broadening the reach of participation, and creating a personalized environment for introspection. Meanwhile, the face-to-face workshops still played a crucial role in bolstering community ties and trust. This research contributes to understanding chatbots’ strengths and weaknesses in participatory processes, both within spatial design and beyond. In addition, it informs future explorations of participatory processes that span different spatial-temporal configurations.

Keywords

Artificial intelligence chatbot community participation large language models natural language processing participatory design

Introduction

Community participation in architectural and urban design processes is critical to ensuring that built environments reflect stakeholders’ lived experiences, values, preferences, and needs.^1-6 To date, participatory processes for architectural and urban design most often rely on group-based, face-to-face community meetings or design workshops as the primary, often the sole activity of participation.^7-9 While useful in many regards, such as fostering in-the-moment idea deliberation,^10,11 strengthening community ties,^12,13 evoking activism and advocacy,¹² and empowering stakeholders as co-creators and partners with equal says,¹⁴ these meetings and workshops are time- and resource-intensive to facilitate.^15-17

These practical constraints often lead to limited engagement with stakeholders that are systematically marginalized¹² or geographically distributed.^18,19 Indeed, the demographics of those who participate in such workshops are known to be skewed toward those with higher socioeconomic status.²⁰ Furthermore, in low-resource contexts, participation is often outright obsolete.¹⁶ Working at an urban scale further compounds this challenge of inclusion,^21-23 given the challenges to sustain meaningful participation over time and space with a diverse set of stakeholders.

To work around this challenge, researchers across architecture, urban planning, human-computer interaction (HCI), and participatory design (PD) have explored alternative ways to involve stakeholders in the design and planning processes of built environments leveraging information and communication technologies (ICTs). Prior research has explored how ICTs can strengthen community participation at urban scale.^21,24-26 Some experimented with replicating the workshop experience using videoconferencing and collaborative sketching tools.^10,11,27-30 Others embraced the Asynchronous Remote Community (ARC) method,³¹ facilitating online discussions through social media platforms, message boards, and group chats.^26,32-35 Despite the breakthroughs that came with these endeavors, challenges remain when it comes to balancing between fostering meaningful participation and scaling up the reach of participation.

Recent advancements in Large Language Models (LLMs) present the maturing capabilities for conversational agents, also known as chatbots, to automate flexible, fluent dialogues at the level of complexity that is commensurate with design conversations for spatial design and planning. This presents the potential for conversational agents to scale up community participation while ensuring that such participation is meaningful, genuine, and empowering. Recent studies have shown the effectiveness of LLM-powered conversational agents in eliciting individual preferences and facilitating the exchange of opinions in design conversations within controlled environments.^36–41

This paper builds on these early successes and explores the use of an LLM-powered chatbot in a real-world architectural design process for a mixed-use building that incorporates community spaces. We reported our experience with a hybrid approach to community participation, in which we integrated (1) one-on-one, remote chatbot engagements, and (2) group-based, face-to-face workshops. We devised our hybrid approach with an eye on scaling up the reach of participation while concurrently building on the known strengths of design workshops.

Our hybrid approach presents a promising direction for community participation in spatial design and planning processes. Our chatbot system was effective in creating a personalized environment for introspection, allowing stakeholders to reflect on their ideas before sharing them with others. Additionally, it elicited differing viewpoints from marginalized populations. On the other hand, our design workshops played a critical role as informal social spaces that bolstered community ties and trust.

The contribution of the paper is twofold. First, this paper describes the strengths and weaknesses of chatbots as an emerging method of community participation through a real-world architectural design project. As LLMs offer the maturing capabilities to generate flexible, fluent dialogues in multiple languages,⁴² this paper serves as a valuable reference for future research that seeks to incorporate chatbots in design processes, both within spatial design and beyond. Second, through the hybrid approach, we examined how the two methods of participation might amplify the strengths and mitigate the shortcomings of one another. Insights from our research can motivate future explorations of design processes that span different spatial-temporal configurations.

The paper is structured as follows: We first review the participatory turn in spatial design that motivated our inquiry. Then, we expounded on our case study and design context. We then presented our observations from the design process. Lastly, we reflected on our experience, discussing design implications and future applications.

Participatory turn in spatial design

Since Henry Sanoff advocated for the pursuit of deliberative democracy in the design of built environments, the past two decades saw a “participatory turn”⁴³ in spatial design where practitioners aim to “discover how to make it possible for people to be involved in shaping and managing their environment.”⁴⁴ From community design,^44,45 community architecture,⁴⁶ to community participation,⁴⁷ researchers and practitioners have been iterating through how to operationalize participation in spatial design practices. These approaches are united by the idea that participatory approaches to spatial design “seek a citizen voice capable of recognizing other group’s interests, appreciating the need for tradeoffs, and generating a sense of common ownership.”⁴⁸ Participatory approaches to spatial design have since expanded beyond traditions in deliberative democracy^6,44 to engage with scholarship in Participatory Action Research (PAR),⁴⁴ Participatory Design (PD),⁴⁹ and co-design.^2,3,50 Across these traditions, the consensus holds that meaningful participation not only gives stakeholders access to the decision-making process, but further empowers them to affect outcomes.^4,6,51

Sherry Arnstein⁶ argues that participation is valuable to the extent that it “is the redistribution of power that enables the have-not citizens …to be deliberately included in the future.” In this vein, “genuine” participation⁵² reflects a fundamental shift from seeing the stakeholders merely as informants to viewing them as partners in the design process.⁴⁹ This equalizing of power is critical in empowering stakeholders to take part in the re-envisioning of their built environments.⁴⁴ In recent years, there has been growing emphasis on this conception of empowerment, particularly in regard to envisioning inclusive, sustainable, and resilient futures of our environments.^53–55

To date, the most common approach to facilitating stakeholder participation in spatial design is hosting face-to-face community meetings and design workshops. We follow Christina Harrington and colleagues¹² in defining design workshop as “a spatially situated and temporally bounded coming together of participant groups and researchers to envision new design futures, which employ particular materials, tools, and goals.” These workshops are synchronous group activities where stakeholders simultaneously participate in the design process via verbal communication and in-situ engagement with physical or digital artifacts. Prior research shows that such workshops are effective in empowering stakeholders as design partners with equal says as the facilitators,¹⁴ fostering in-the-moment idea deliberation,^10,11 strengthening community ties^12,13 and in evoking thoughts and ideas towards social changes.¹²

Researchers and practitioners favored the design workshop format for various reasons. For instance, these workshops have proven to be effective in fostering in-the-moment idea deliberation,^10,11 strengthening community ties,^12,13 evoking activism and advocacy,¹² and empowering stakeholders as co-creators and partners with equal says.¹⁴ Despite their strengths on these fronts, face-to-face workshops come with constraints that may be inadequate in realizing the participatory ideals that researchers and practitioners set out to achieve. On the participants’ end, Archon Fung⁵¹ highlights that the vast majority of those who attend such events “do not put forward their own views at all” and instead participate as “spectators.” Even when stakeholders do express their views, design workshops as a culturally constructed practice privileges the ideals of innovation as defined by the scholarly applications of design thinking, leaving limited room for a pluralistic understanding of design and imagination as defined by different stakeholder groups.^12,56,57 For instance, “engaging in design processes that promote ‘blue-sky’ ideas (or ideation without constraints) may exacerbate inequities by leading to infeasible solutions that ultimately frustrate underserved individuals.”¹² As a result, systematically marginalized groups are less likely to meaningfully engage in design conversations even when present in these settings.^{12,56,58–61} Furthermore, partaking in such workshops requires participants to be absent from their day-to-day commitments (e.g., work, family duties, leisure) and dedicate hours to the commutes and design activities.^{15,16,56,62,63} This prerequisite to attendance often makes participation in workshops a luxury that is only accessible to those who can afford to set aside several hours at a time for such activities. As a result, those with the privilege to participate and innovate in ways that align with the ethos of design facilitators get to define the success of such activities, which further marginalizes the values and needs of underserved communities.⁵⁹

On the organizers’ end, design workshops are time- and resource-intensive to facilitate.^15,16 Such practical constraints of workshops lead to several downstream effects on overall design processes. First, such workshops largely take place as one-off activities in the early stage of the design process.^64–67 As a result, stakeholders primarily only “participate” when the design concepts are in the form of written design briefs or rudimentary two-dimensional sketches, and their involvement becomes obsolete once the design ideas become more complex.⁶⁸ Second, these workshops are primarily single-site activities that are not accessible to those who are geographically dispersed^14,69, thus sidelining their voices in the design process. Lastly, design workshops are often absent in low-resource communities when facilitators cannot afford to host such workshops.^16,70,71 These constraints of design workshops pose challenges to fostering meaningful participation at an urban scale.^21-23

Scaling up via new technologies

Two decades ago, Neumann and Star⁷² questioned to what extent stakeholder participation could be realized in large-scale design projects. With projects that are temporally and geographically distributed, and with stakeholder groups that are heterogeneous, how can design and planning professionals meaningfully involve stakeholders in the design process? Many earlier attempts to scale up design projects have led to failures or limited successes.^73,74

Indeed, scaling up and careful human interventions are often conceived of as naturally opposed to each other.^75–77 This raises questions about the extent to which meaningful participation is possible at scale.^{8,22,53,64,78–84} In other words, is it possible to extend and sustain the reach of participation without giving up on the core tenets of participation?

Against this backdrop, design and planning professionals called for a renewed understanding of participation that adjusts the goals from “projecting” to “infrastructuring.”^85–88 In this view, the designers' responsibility extends beyond the timescale of a single project;, re-orienting toward a continuous “process of becoming” within an ecosystem.^85–88 This reflects the increasing awareness around the importance of considering longer temporal horizons⁸⁷ with a realignment from a project focus to use in an unspecified future. The design project as the prevalent and taken-for-granted temporal unit of limited duration has led to a change in focus, from product to process, creating opportunities for future design-in-use,^62,85 extending design towards more open-ended, long-term processes.^86,89 Infrastructuring, then, is the work of creating sociotechnical resources that intentionally enable adoption and appropriation beyond the initial scope of the design, a process that might include participants not present during the initial design engagement.⁸⁸

In the spirit of infrastructuring, researchers have explored alternative ways to involve stakeholders in the design and planning processes. Some experimented with hosting design workshops online using videoconferencing and collaborative sketching tools.^10,11,27-30 These tools hold great promise for stakeholders to participate at their locative convenience while still supporting synchronous idea deliberation and relationship-building as in-person workshops. Furthermore, these workshops also supported participants in becoming comfortable building on the ideas of other participants gradually.

However, online workshops still present several constraints. First, on the participants’ end, they still require a significant time commitment and presume particular conceptions of innovation, posing practical constraints on how these workshops can include diverse voices.¹⁰ Second, they pose new barriers to participation as meaningful engagement in such workshops requires a certain level of digital and technical literacy on the participants’ end.²⁸ Third, technical constraints of existing tools can be disruptive to idea deliberation and relationship-building. Lee et al. noted the lack of visual cues for facilitators and participants to interpret the emotions of one another.¹⁰ Furthermore, Harrington and Dillahunt found that participants showed resistance to long periods of Zoom sessions with their cameras on.¹¹

Meanwhile, other professionals have attempted to address these challenges through the ARC method.³¹ The ARC method facilitates online, web-based discussions through social media platforms, message boards, and group chats that support stakeholders across geographic and resource differences to participate at their convenience.^26,32-35 Unlike virtual or in-person workshops that tend to be one-off activities, these messaging tools naturally support stakeholders to participate over time iteratively. Furthermore, messaging-tool-based participation took up far less time (a few minutes) than design workshops (usually lasting for an hour or more). However, the inability to engage participants in interactive activities at the same level as synchronous workshops is a known limitation of the ARC method.^10,11

Yet others explored alternative approaches to stakeholder engagement, such as crowdsourcing data,^25,90 creative crowdsourcing,^67,68 participatory GIS,^91–93 sensing technologies and the Internet of Things (IoTs),^94,95 blockchain technologies,⁹⁶ and serious games,^97,98 among others. Despite the breakthroughs that came with these endeavors, challenges remain when it comes to balancing between fostering meaningful participation and scaling up the reach of participation.

Conversational agents as the new frontier

The past decade saw the emergent applications of Artificial Intelligence (AI) in generating ideas for spatial design.^99–105 In particular, researchers have been exploring the potential of graph-based machine learning systems via techniques like deep learning and neural networks. Most recently, many have developed novel technologies based on Generative Adversarial Networks (GANs) that can automate the process of graphical generation.^106,107 This body of work primarily focuses on how AI can help generate new conceptual floor plans or layout designs.

Moving forward, we advocate for re-orienting our emphasis toward the potential for AI to not only foster — but scale up design conversation.^65,108–112 To this end, we ask — how can AI scale up participation while augmenting — rather than diminishing — the dialectical, communicative, and even agnostic nature of participation. After all, participation is a process where participants “clarify, explain, interpret, assess, argue, and engage in iterative levels of reflection and critique”.¹¹² If we look beyond traditional liberal conceptions of democracy where participation is exemplified by rational negotiation among interests toward consensus,^113,114 participation is also defined by the arena where people confront our differences and adversaries.^115–117

It is in this vein that we turn our attention to conversational agents in fostering the space for design conversations — whether they are consensus-reaching or agnostic by nature. First, chatbots are a cost-efficient technology that assists communication among people across geographic and resource differences and through the design process.^118,119 Second, compared to videoconferencing and other synchronous tools, chatbots have the distinctive advantages of temporal flexibility and scalability¹²⁰: participants can engage in design conversations at times of their convenience, and designers can easily engage with a large number of participants via one chatbot. Third, in contrast to the various messaging tools, chatbots can carry out context-aware and even open-ended conversations with different participants to foster idea deliberation.^121,122

Recent studies have demonstrated successes in utilizing LLM-powered conversational agents to elicit individual preferences and facilitate the exchange of opinions for design discussions in controlled environments.^36-40 This paper builds on these early successes and explores the use of an LLM-powered chatbot in a real-world architectural design process for a mixed-use building that incorporates community spaces.

Methodology

In this paper, we reflected on our experience leveraging an LLM-based chatbot as part of the design process for a multi-purpose community center that will be built in the Galilee region near the Israel-Lebanon border. The paper presents one part of a larger project that leveraged both crowdsourcing and language technologies to support design interactions on an urban scale. This paper zooms in on the potential of language technologies in fostering stakeholder participation during the design brief and concept design stages. (See The Royal Institute of British Architects “plan of work” that outlines the sequential stages of the architectural design workflow.¹²³) Insights on the crowdsourcing system architecture have been published in another paper⁶⁸.

Local contexts

The community center is situated in a region characterized by an ongoing conflict between Israel and neighboring countries, as well as between the Jewish and Arab communities.^124,125 Figure 1 illustrates the geographic context of this project. To design the community center in ways that serve these different stakeholders, the inclusion of diverse and heterogeneous voices across geographic, political, resource, and language differences is crucial to its success.

Figure 1.

Geographic context of the Tzahar Innovation, Culture, and Education Center.

In-person workshops alone would be inadequate for achieving this goal. First, the region’s geopolitical conflicts pose challenges to co-located participation across stakeholders who hold contentious opinions. Second, the project concerns a wide array of stakeholder groups with varying socioeconomic status (SES), ranging from architects, government officials, community leaders, business owners, to blue-collar workers. Third, the multilingual context of the region – where stakeholders have varying proficiency in Hebrew, Arabic, and English – presents unique challenges to the moderation of synchronous conversations. Fourth, the workshop format is not accessible to many relevant stakeholder groups. For instance, married women from these Arab communities are less likely to attend such workshops on their own.

Designing the design process

Taken together, the case of the community center presents concrete needs for practitioners to explore alternative methods for stakeholder participation. Nonetheless, we acknowledge the strengths of design workshops. Instead of a complete abandonment of face-to-face workshops, we innovated a hybrid approach to PD that integrated three group-based, face-to-face workshops and four sessions of one-on-one, remote chatbot engagement (see Figure 2).

Figure 2.

Participatory design process for the community center.

To examine if and how chatbots may sustain participation throughout the design process, we planned four sessions of chatbot interactions that mapped onto the transition points between different forms of design artifacts, from design briefs, co-design sketches, 2D sketches, to 3D computer-aided designs (CADs).⁶⁸

In addition, to facilitate the transition between the different stages of the architectural design process (briefing stage, concept design stage, and onward), a professional group facilitator led the face-to-face workshops at the beginning and end of these stages to consolidate the design discussions.

To foster flexible modes of participation, we had no prerequisite for any of these points of engagement. The stakeholders had the flexibility to participate at any point, in any form, and as frequently as they preferred. Stakeholders may participate for the first time mid-way through the PD process, or they may skip a few sessions in between their sessions of participation.

With approval from our Institutional Review Board, we recruited our participants through our project website. For those who signed up and expressed interest in participating in the process, we sent out periodic email updates on the project status, as well as invitations to our workshops and chatbot engagement.

We circulated the website within local online communities and advertised it on Facebook, reaching an estimated audience of over 15,000 individuals. We also reached out to key opinion leaders in the local communities, including entrepreneurs, artists, and educators, asking them to disseminate information about our project.

Designing the chatbot

We designed the chatbot system, ArchitectureBot, to support the stakeholders in developing their design ideas and critiquing ongoing design efforts. We developed ArchitectureBot using GPT-3 DaVinci⁴² as our foundational model since it could carry fluent open-ended conversations. We utilized Google Translate to support the chatbot in English, Hebrew, and Arabic. The stakeholders engaged with the system through either text-based or verbal exchanges based on their preferences. Interactions with the chatbot were facilitated through an HTML client that was mobile-friendly and resembled an instant messenger (see Figure 3 that depicts the conversation flow.).

Figure 3.

Conversation flow of ArchitectureBot (from left to right): 1) Ideation/design review - for the first session, the chatbot prompts the stakeholder to express their opinions and aspirations for the project. For the subsequent sessions, the chatbot prompts the stakeholder to review the latest design progress. 2) Thought elaboration - once the stakeholder selected their preferred designs, the chatbot prompts them to elaborate on how to improve the design. 3) Identifying requirements - the chatbot uses the conversations to identify design requirements, which are exported for the architects’ reference for the subsequent design iteration.

For the first session of chatbot engagement, the chatbot would prompt the stakeholder to express their preferences and aspirations for the community center. The chatbot would respond to each participant’s message with follow-up questions or suggestions (see Appendix for detailed prompts). After the conclusion of all engagements, the chatbot would aggregate all participants’ inputs into a written summary, which served as the starting point for architects to visualize these design requirements into concept sketches.

For the subsequent sessions of chatbot engagement, the chatbot would prompt the stakeholder to review and provide feedback on the latest design ideas (see Appendix for detailed prompts). These designs were in various formats across the design process, from napkin sketches, 3D models, to CAD drawings. Once the stakeholder had reviewed all the latest designs, the chatbot would prompt the stakeholder to select specific designs and elaborate on how to improve the designs. For each session of engagement, the chatbot would process these conversations to identify and export design requirements for the architects’ reference for the next design iteration.

Workshops

The project involved three in-person workshops, as illustrated in Figure 2. These workshops were led by a professional facilitator and supported by three architecture graduate students.

The first workshop aimed to discuss the project’s needs and create a design brief. It began with an ice-breaker activity to introduce participants to each other. Following this, participants were divided into four groups. Each group focused on a different theme: aesthetics, function, sustainability, and community. The groups were then asked to propose design ideas related to their theme. After the group discussions, the groups presented their ideas to all the participants. The workshop was video recorded and later transcribed for analysis. All design ideas from the transcription were manually extracted. These ideas were then combined with the results of interactions with a chatbot. These ideas were compiled into a design brief using our chatbot system. This document was shared with the developer and chatbot participants through Google Docs for their review and comments.

The second workshop centered around a co-design activity in which the participants expressed their design ideas using napkin sketches. Three architecture students helped the participants to express their ideas graphically. The workshop began with a presentation about techniques to express conceptual design ideas. The produced sketches were collected, scanned, and uploaded to the crowdsourcing system for further evaluation and integration.

The final workshop came in two parts. The first half entailed a comprehensive presentation of all proposed designs and a discussion among the participants on their preferred designs. The second half entailed a reflexive discussion on how the participants felt about the overall experience of participating in the design process.

Reflecting on the design process

We reflected on our experiences and triangulated our findings across various data sources.¹²⁶ These include the logs of all conversations our participants had with our chatbot, the transcripts of the discussions that took place during the face-to-face workshops, the field notes that capture the human interactions during these workshops, as well as the video recordings of these workshops.

Findings

The community design process took place from April to June 2022, involving 46 participants from a pool of 118 registrants who expressed interest via our website. The participants represented a broad age spectrum, ranging from 20 to 78 years, with a mean age of 39.6 (std. 13.95). Gender-wise, more women (56.8%) than men (40.7%) participated in the process. A total of 1120 chat messages were recorded and distributed across 141 conversations with 29 unique participants.

Below, we highlight four key findings from the process: First, our case study illustrates the usefulness of chatbots in sustaining participation over time. Second, our experience confirmed LLM-powered chatbots’ capacity to diversify participation across geographic, resource, and language constraints. Third, we found that our chatbot presented the strength of fostering a personalized, private space for introspection, a feature absent from communal participation forms. Fourth, we noted that our design workshops shined in fostering social connections among the participants, while our chatbot was unable to do so. Consequently, we argue that our hybrid method brought forth the strength of both.

Sustaining participation over time

Throughout the design process, we saw a considerable drop in the number of participants in our design workshops. 18 participants attended the initial workshop, which led to the identification of 44 design requirements. 11 participants took part in the second (co-design) workshop and produced 10 sketches. Four participants joined the third workshop for design presentations and reflexive discussions. Notably, the final workshop attendees were a subset of the initial workshop, indicating declining participation over time without new participants.

By contrast, our chatbot system was somewhat successful in sustaining participation over time. Our results show that most of those signed up to interact with the chatbot remained engaged throughout the design process. The attrition rate of chatbot interactions is low compared to the workshops. In the first week, seven participants engaged with the chatbot to provocative design ideas, resulting in 36 design requirements. In the subsequent weeks, the chatbot facilitated the iterative review of concept design. 32 participants engaged in the first round, generating 36 ideas for improvement. However, we note that some participants did not finalize their conversations. Next, 18 participants engaged in the second round, leading to 43 ideas for improvement. Lastly, 16 participants engaged in the last round, resulting in 50 ideas for improvement.

With the integration of the chatbot system, the temporal extension of the participation period allowed stakeholders to become more familiar with the design goals and objectives. It also allowed stakeholders to provide feedback and input throughout different stages of the design process. Compared to traditional workshops that largely serve as one-off participation in the early phase of the design process, the advent of LLM-powered chatbots presents a technical breakthrough in fostering long-term participation. Yet, it is crucial to note that our case only involved 46 participants over a span of 3 months. It remains to be further explored if this capacity to sustain participation still holds if we further scale up the design project.

Extending the reach of participation

Our use of the chatbot system was able to extend the reach of our participant pool. Among the 29 people who participated via chatbot, 18 of those are unique participants who did not participate in the workshop series.

In particular, as a result of the use of the chatbot system, the project saw considerable geographic diversity among the participants. While only 23% of the workshop participants resided in areas outside of the immediate neighborhood where the community center would be built, the number increased to 54% among chatbot participants. This indicates how our chatbot was useful in reaching those who reside in areas further away from where the workshops were held —- those who would have otherwise not participated in our process.

Meanwhile, the project saw relative success in engaging the traditionally under-represented Arabic minority through the chatbot system, albeit only to a limited extent. Seven individuals from this community enrolled and subsequently participated in the discussion for our project through our online platform. Two of these participants identified as women, and they participated in all the chatbot interactions of the project. While these numbers are small, we believe this observation is noteworthy considering the cultural norms within many traditional Arab communities, where women are often unlikely to participate in physical meetings with unfamiliar individuals. The inclusion of Arab women via our chatbot suggests the potential for this technology to serve as a more accessible method of participation among traditionally under-represented groups. These are people who would very likely not have participated in our process otherwise. This outcome illustrated that our chatbot system extended the design conversations beyond the resourceful stakeholders.

By contrast, all but one of the participants of the workshops come from Jewish localities and present high SES as entrepreneurs, government officials, or community leaders. Only one member from the Arabic communities, a male community leader, ever participated in our design workshops. Yet, his involvement was only made possible through a personal invitation extended by the project coordinator. Furthermore, he only participated in the first workshop; our subsequent workshops failed to attract and include any Arab participants.

Personalized space for introspection using chatbots

Our design process highlighted dyadic chatbots’ unique strength in fostering a personalized, private space for introspection. The chatbot did so in two ways – by (1) inquiring stakeholders to elaborate on their thoughts and by (2) suggesting new ideas to stakeholders. To illustrate how the chatbot supported this process of introspection, we drew on excerpts from conversations that took place between the chatbot and participants.

The first example illustrates how the chatbot prompted the participants to develop their thoughts further by conversing with it. The participant expanded their idea from the theme of accessible space to shading elements like fins on the main road to the building.

1. Inquiring for further details:

CHATBOT: “What do you like about it? How should we improve it?”

PARTICIPANT: “The use of the structure should be appropriate.”

CHATBOT: “Do you have any specific ideas in mind?.”

PARTICIPANT: “For all ages and accessible. It is important that access to the building from the main road be accessible”

CHATBOT: “Yes, that is a good point. What else?”

PARTICIPANT: “Ramps for accessible spaces”

CHATBOT: “Yes, the design does need to be more accessible. Are there any other specific features you would like to see in the design?”

PARTICIPANT: “Shading elements like fins on [the] East and West side and roofs on [the] North and South”

The second example illustrates how the chatbot prompted the participants to consider ideas different from but adjacent to what they initially expressed. The participant expanded on their thoughts to consider both surface and underground parking spaces.

2. Suggesting new ideas:

PARTICIPANT: “Where do people park?”

CHATBOT: “We’ll need to include parking in the design. Would you like to see surface parking or underground parking?”

PARTICIPANT: “Underground parking is too expensive.”

CHATBOT: “Okay, we’ll make sure to include surface parking in the design. Is there anything else you would like to discuss about the design?”

In both cases, the maturing of ideas was made possible by the chatbot’s capability to build on open-ended, flexible inquiries to personalize each person’s interaction. This personalized nature of chatbot interactions contributed to an uninterrupted experience of idea development. The entire experience of interacting with the chatbot was tailored for individual stakeholders based on their interests.

By contrast, during the face-to-face workshops, we observed that the presence of other stakeholders often led to digressions in the discussion topics. For individual stakeholders, this can be disruptive to their idea development processes. The following dialogue about having a maker space in the community center during the brainstorming session illustrates how such digressions took place:

3. Changing topics in group discussions:

PARTICIPANT A: “For workshops or some forms of maker space…something that will force people to exchange ideas and go home with [new thoughts], whether they want to or not. Say if they pass by a corridor, experience works of art, or see their neighbors, these can all inspire their minds.”

PARTICIPANT B: “A meeting of cultures, which I think is a very important part of this story.”

PARTICIPANT C: “It is important to mobilize young people from the region to discuss a societal problem…to find a solution together with the young people in a room called the “Maker Room,” with all the facilities they might need. For example, I [Participant C spoke about a specific project that we cannot mention for privacy reasons].”

As illustrated by the conversation above, the discussion did not help any of these participants to develop their thoughts further. Rather, the participants leveraged the workshop to fulfill their intention of agenda-setting and self-expression. We argue that had the same discussion taken place over the chatbot, the chatbot would have prompted each participant to further explore these ideas instead of digressing to another topic prematurely.

With chatbot interactions, the private space for introspection was particularly helpful for eliciting diverse opinions and deliberating conflicting opinions. During the workshops, the participants rarely voiced, confronted, or resolved conflicting opinions. Most interactions during the workshops were affirmative or supportive of other participants’ ideas. By contrast, we found multiple instances where the participants expressed dissenting opinions when interacting with the chatbot in private.

We attributed this contrast to two reasons:

1. Personalized dialogues afford local conventions of innovative approaches. With personalized dialogues, the process of idea development was tailored for the sensemaking process of individual participants. In this case, the participants did not have to conform to the ideals of innovation as defined by the facilitators. The chatbot empowered them to articulate their preferences, needs, and values in alignment with their local ethos.

2. One-on-one human-chatbot interactions decoupled the participants’ identities and their opinions. When expressing their opinions in the design workshops, participants had to inevitably make their identities known to the other attendees, and their identities would be inextricably linked to their opinions. By contrast, when participating through the chatbot, participants could influence the design process through their opinions without “outing” themselves to the other stakeholders. This quasi-anonymous nature of one-on-one human-chatbot interactions contributed to a more diverse corpus of ideas that often included conflicting opinions.

The findings suggest that chatbot-mediated interactions foster a personalized space for introspection and idea development, particularly beneficial for eliciting diverse opinions and deliberating conflicting opinions. Compared to the workshop discussion dynamics, participants could express their ideas freely without being criticized or interrupted. The results demonstrate how dyadic chatbots can be leveraged to create more inclusive spaces where stakeholders feel comfortable expressing their ideas without fear of judgment or criticism from other participants.

Fostering social relationships

Our case study also pointed to a critical limitation of one-on-one, remote chatbot engagement as a method of participation: Despite the capacity to automate human-like dialogues, our chatbot could not foster social connections among the participants. While we previously highlighted the decoupling of identities and ideas as the dyadic chatbot’s strength in eliciting conflicting opinions from the stakeholders, such decoupling also took away the social benefits of workshops by knowing other stakeholders’ identities and making one’s identities known to others.

By contrast, how did the workshops afford the participants to build social connections with others and a public presence in the community? While dyadic chatbot engagement fostered the personalized space for introspection, workshops fostered a collective experience of co-presence. This experience created an opportune setting for people to bond with one another as social and public beings. In this sense, in-person workshops served as the settings for informal, spontaneous social encounters with other stakeholders.

From the workshops, we observed that informal, spontaneous conversations before and after formal discussions were key to forming such social connections and public presence. After the workshops, some participants would approach and talk to people they had just interacted with during the formal discussions. In particular, the workshop attendees expressed that they were especially enthusiastic about connecting with people from stakeholder groups that are different from their own. One of them highlighted that they were especially interested in connecting with people “they would have never met if not for the [workshop].” Such spontaneous encounters were only possible in a collective experience like design workshops where all stakeholders were invited to congregate in the same environment.

We also acknowledged that the success of our workshops in fostering social bonds was limited to the particular type of people who would attend these workshops in the first place, meaning those who come from dominant groups and those who have higher SES status. As the design workshops attracted a rather homogeneous corpus of attendees, with this method, we had no success in fostering social connections across stakeholder groups with geographic and resource differences.

To this end, our findings show that combining group-based, face-to-face workshops and one-on-one, remote chatbot engagement in a hybrid design workflow could realize the strength of both. The chatbot system was effective in sustaining participation across the design process. In addition, it was effective in extending the design conversations to those who are geographically dispersed or in low-resource settings. Meanwhile, the design workshops were useful in fostering social connections among stakeholders through the workshops. Such outcomes demonstrated that the current hybrid workflow could realize the strengths of a dyadic chatbot combined with the strengths of face-to-face workshops as a method of participation.

Implications and future directions

Drawing from our experience, we highlighted four lines of future directions for practitioners and researchers to further explore. The provocations below serve to inspire practitioners and researchers as they consider incorporating chatbots as part of their PD processes, or as they look into new spatial-temporal configurations of design processes more broadly. After all, our approach presents merely one form of a hybrid approach to community participation in design and planning processes, and we only explored one form of a chatbot-infused hybrid process.

Calibrating chatbots toward introspection

Our case study demonstrated the effectiveness of chatbot engagement in fostering a private space for introspection. However, we did not explore how the design of the chatbot itself may influence its effectiveness. Notably, our chatbot did not maintain any memory to personalize the interactions for individual stakeholders across different sessions of engagement. In addition, our chatbot did not prompt the stakeholders to reflect on the opinions of others during the conversations. Future research could investigate how these factors may further support stakeholders’ acts of introspection throughout the design process.

Bridging spaces for socialization and introspection

Our hybrid processes highlighted two spaces that were critical to the stakeholders’ satisfactory and meaningful participation – a public sphere for socialization and a private space for introspection. The group-based, face-to-face design workshops afforded the former, and the one-on-one, chatbot engagement afforded the latter. This highlights the strength of our workflow in concurrently fostering these two spaces by converging different spatial-temporal configurations – between synchronous and asynchronous, between face-to-face and remote, and between public and private. Noteworthily, our combination of the two methods is merely one example that can simultaneously foster the two spaces. We encourage researchers to explore alternative design workflows that can also foster these two spaces in a concurrent fashion. In our case study, we did not explore how we could best bridge the two spaces in a single design workflow. For instance, how may the frequency and sequence of design workshops and chatbot engagement affect the effectiveness of this combination of methods? Are there other methods of engagement that can also be used in conjunction? We call for future research to further this line of inquiry.

From chatbots to social bots?

From our case study, we learned that our remote engagement attracted a more diverse range of stakeholders compared to their in-person counterparts. As such, we encourage future research to explore the potential of fostering informal spaces for socialization as part of their online interactions. While we deployed our online chatbot in a dyadic manner, future researchers can consider re-configuring the context of chatbot interactions to make room for social bonding. For instance, researchers can deploy conversational agents as moderators to foster many-to-one, one-to-many, or many-to-many human-bot conversations. This means looking beyond the function of chatbots as merely for information exchanges, but to see chatbots as a form of a social bot with the communicative capacity to foster social and communal ties.

Sociotechnical re-orientation for broadening participation

Through the use of our chatbot system, we saw successes in sustaining participation over time and extending the geographic scope of participation. By contrast, we only saw rather limited successes in engaging with the most marginalized communities in the region — the Arab communities. While we were able to include a few Arab people in our design process via our chatbot, our case illustrated the limitation of a purely technical approach to community engagement. Similarly, while we saw successes in fostering social connections through informal, spontaneous encounters at in-person design workshops, we were not able to afford the same experience for participants from marginalized backgrounds, since they did not attend these workshops in the first place.

To this end, we call for a sociotechnical re-orientation of the potential for chatbots to foster community participation. This means paying equal attention to the social infrastructures that situate these technologies in the first place. For instance, in the case of chatbot engagement, a sociotechnical approach to extending participation could mean demonstrating the usefulness of chatbot interactions at the local community level, such as by presenting project ideas and chatbot interactions at local community centers. Similarly, in the case of design workshops, one can consider bringing these workshops to locations that more conveniently serve communities that are traditionally under-served in design processes.

Conclusions

This paper presents a case study of a hybrid workflow for community participation in a real-world architectural design project. We combined in-person, group-based workshops with remote, one-on-one chatbot engagements to support the design ideation of a multi-purpose community center. Our findings illustrate three strengths of chatbots as a proxy of participation. We also noted the limitation of such a method, highlighting how the integration of workshops in our hybrid process was complementary to chatbot engagements.

First, we noted the usefulness of chatbots in sustaining participation over time. With the integration of the chatbot system, the temporal extension of the participation period allowed stakeholders to become more familiar with the design goals and objectives. It also allowed stakeholders to provide feedback and input throughout different stages of the design process. Compared to traditional workshops that largely serve as one-off participation in the early phase of the design process, the advent of LLM-powered chatbots presents a technical breakthrough in fostering long-term participation. Yet, it is crucial to note that our case only involved 46 participants over a span of 3 months. It remains to be further explored if this capacity to sustain participation still holds if we further scale up the design project.

Second, we highlighted the potential for chatbots to extend participation beyond resourceful, co-located stakeholders. We noted that our chatbot was helpful in extending the scope of participation across geographic constraints, particularly those that reside in areas further away from where the workshops were held. Nonetheless, we only saw limited successes in engaging with the most marginalized communities in the region — the Arab communities. While we were able to include a few Arab people in our design process via our chatbot, we note that the extent to which chatbots can extend the reach of participation remains to be critically examined.

Third, we found that our chatbot was effective in fostering a personalized space for introspection and idea development, particularly beneficial for eliciting diverse opinions and deliberating conflicting opinions. The chatbot did so in two ways – by (1) inquiring stakeholders to elaborate on their thoughts and by (2) suggesting new ideas to stakeholders. Compared to the workshop discussion dynamics, participants could express their ideas freely without being criticized or interrupted. The results demonstrate how dyadic chatbots can be leveraged to create more inclusive spaces where stakeholders feel comfortable expressing their ideas without fear of judgment or criticism from other participants.

Lastly, we highlighted one strength of design workshops that our chatbot lacked — providing informal, spontaneous spaces that bolster communal ties. These are spaces that our participants deemed as critical to a participatory process. To this end, our findings show that combining group-based, face-to-face workshops and one-on-one, remote chatbot engagement in a hybrid design workflow could realize the strength of both.

It is crucial to note that our approach presents merely one form of a hybrid approach to community participation in design and planning processes. Furthermore, we only explored one form of a chatbot-infused hybrid process. Given the known challenges that we encountered in our process, namely the difficulty of extending the reach of participation to marginalized communities, we encourage researchers and practitioners to future explore alternative possibilities that build on our empirical insights.

Limitations notwithstanding, this paper makes two contributions. As LLMs offer the maturing capabilities to generate flexible, fluent dialogues in multiple languages, our findings serve as a valuable source of reference for future research that seeks to incorporate chatbots in design processes, both within spatial design and beyond. More broadly, the successes and limitations of our hybrid approach can shed light on future explorations of design processes that span different spatial-temporal configurations.

Yet, we noted that our success was limited, and we highlighted that the effectiveness of the technology remains to be further explored in future studies and implementations.

Footnotes

Acknowledgement

The research team would like to thank Dan Nissimyan for providing us with the opportunity, generous funding, and inspiration to make this project happen. We would also like to thank our research assistants from Technion - Israel Institute of Technology for their work, incuding Roni Hillel and Tamar Levinger from the Material Topology Research Lab (MTRL), as well as Dasha Sobotine. We would also like to thank the team at Valiant Investment for their support, particularly Eylon Yadin, Omer Aviram, and Mike Schnall.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Valiant Investments LTD and the Israeli Ministry of Innovation, Science, and Technology (grant no. 2032704) .

ORCID iDs

Stephen Yang

Jonathan Dortheimer

Appendix

References

Jones

Petrescu

Till

. Architecture and participation. London, UK: Routledge, 2013.

Sanders

EBN

Stappers

. Co-creation and the new landscapes of design. CoDesign 2008; 4(1): 5–18.

Steen

. Co-design as a process of joint inquiry and imagination. Des Issues 2013; 29(2): 16–28.

Polletta

. Freedom is an endless meeting. Chicago, IL: University of Chicago Press, 2012.

Putnam

. Bowling alone: America’s declining social capital. Jod 1995; 6(1): 65–78.

Arnstein

. A ladder of citizen participation. J Am Inst Plan 1969; 35(4): 216–224.

Halskov

Hansen

. The diversity of participatory design research practice at pdc 2002–2012. Int J Hum Comput Stud 2015; 74: 81–92.

Smith

Bossen

Kanstrup

. Participatory design in an era of participation. Oxfordshire, UK: Taylor & Francis, 2017.

Sanders

EBN

Stappers

. Probes, toolkits and prototypes: three approaches to making in codesigning. CoDesign 2014; 10(1): 5–14.

10.

Lee

Roldan

Zhu

et al. The show must go on: a conceptual model of conducting synchronous participatory design with children online. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, New York, NY, USA, 7 May 2021, pp. 1–16.

11.

Harrington

Dillahunt

. Eliciting tech futures among black young adults: a case study of remote speculative co-design. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. New York, NY, USA, 7 May 2021.

12.

Harrington

Erete

Piper

. Deconstructing community-based collaborative design: towards more equitable participatory design engagements. Proc ACM Hum Comput Interact 2019; 3(CSCW): 1–25.

13.

Hope

D’Ignazio

Hoy

, et al. Hackathons as participatory design: iterating feminist utopias. In: Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, Scotland, UK, 4–9 May 2019.

14.

Duarte

AMB

Brendel

Degbelo

, et al. Participatory design and participatory research: an hci case study with young forced migrants. ACM Trans Comput Hum Interact 2018; 25(1): 1–39.

15.

Irvin

Stansbury

. Citizen participation in decision making: is it worth the effort? Publ Adm Rev 2004; 64(1): 55–65.

16.

Pilemalm

. Participatory design in emerging civic engagement initiatives in the new public sector: applying pd concepts in resource-scarce organizations. ACM Trans Comput Hum Interact 2018; 25(1): 1–26.

17.

Simonsen

Hertzum

. Participative design and the challenges of large-scale systems: extending the iterative pd approach. In: PDC’08: Participatory Design Conference, Bloomington, IN, 1–4 October 2008.

18.

Walsh

Druin

Guha

et al. Layered elaboration: a new technique for co-design with children. In: Proceedings of the SIGCHI conference on human factors in computing systems. New York, NY: ACM, 2014, pp. 1237–1240.

19.

Walsh

Distributed participatory design. In: CHI’11 extended abstracts on human factors in computing systems. New York, NY: ACM, 2011. pp. 1061–1064.

20.

Kavanaugh

Carroll

Rosson

, et al. Participating in civil society: the case of networked communities. Interact Comput 2005; 17(1): 9–33.

21.

Gooch

Barker

Hudson

, et al. Amplifying quiet voices: challenges and opportunities for participatory design at an urban scale. ACM Trans Comput Hum Interact 2018; 25(1): 1–34.

22.

Dalsgaard

. Challenges of participation in large-scale public projects. In: Proceedings of the 11th biennial participatory design conference. New York, NY: ACM, 2010, pp. 21–30.

23.

Corbett

Le Dantec

. The problem of community engagement: disentangling the practices of municipal government. In: Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, Montreal, QC, Canada, 21–26 April 2018.

24.

Asad

Le Dantec

Nielsen

, et al. Creating a sociotechnical api: designing city-scale community engagement. In: Proceedings of the 2017 CHI conference on human factors in computing systems. New York, NY: ACM, 2017, pp. 2295–2306.

25.

Le Dantec

Asad

Misra

, et al. Planning with crowdsourced data: rhetoric and representation in transportation planning. In: Proceedings of the 18th ACM conference on computer supported cooperative work & social computing. New York, NY: ACM, 2015, pp. 1717–1727.

26.

Le Dantec

. Participation and publics: supporting community engagement. In: Proceedings of the SIGCHI conference on human factors in computing systems. New York, NY: ACM, 2014, pp. 1351–1360.

27.

Murai

N Antle

Kitson

et al. Facilitating online distributed critical making: lessons learned. In: FabLearn Europe/MakeEd 2021-an international conference on computing, design and making in education. New York, NY: ACM, 2021, pp. 1–9.

28.

Cajamarca

Herskovic

Lucero

et al. A co-design approach to explore health data representation for older adults in Chile and Ecuador. In: Designing interactive systems conference. New York, NY: ACM, 2022, pp. 1802–1817.

29.

Delcourt

Charmaraman

Durrani

et al. Innovating novel online social spaces with diverse middle school girls: ideation and collaboration in a synchronous virtual design workshop. In: CHI conference on human factors in computing systems. New York, NY: ACM, 2022, pp. 1–16.

30.

Rötkönen

Winschiers-Theophilus

Goagoses

et al. Playing on the globe: facilitating virtual communications between namibian and Finnish learners to co-design an interactive map game. In: Interaction design and children. New York, NY: ACM, 2021, pp. 160–170.

31.

MacLeod

Jelen

Prabhakar

, et al. Asynchronous remote communities (arc) for researching distributed populations. In: PervasiveHealth. New York, NY: Association for Computing Machinery, 2016, pp. 1–8.

32.

Bhattacharya

Liang

Zeng

et al. Engaging teenagers in asynchronous online groups to design for stress management. In: Proceedings of the 18th ACM international conference on interaction design and children. New York, NY: ACM, 2016, pp. 26–37.

33.

Maestre

MacLeod

Connelly

et al. Defining through expansion: conducting asynchronous remote communities (arc) research with stigmatized groups. In: Proceedings of the 2018 CHI conference on human factors in computing systems. New York, NY: ACM, 2018, pp. 1–13.

34.

Danielsson

Naghsh

Gumm

et al. Distributed participatory design. In: CHI’08 extended abstracts on human factors in computing systems. New York, NY: ACM, 2008, pp. 3953–3956.

35.

Prabhakar

Guerra-Reyes

Kleinschmidt

et al. Investigating the suitability of the asynchronous, remote, community-based method for pregnant and new mothers. In: Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems. New York, NY: ACM, 2017, pp. 4924–4934.

36.

Shin

Hedderich

Lucero

et al. Chatbots facilitating consensus-building in asynchronous co-design. In: Proceedings of the 35th Annual ACM Symposium on User Interface Software and Technology. New York, NY: Association for Computing Machinery, 2022, pp. 1–13.

37.

Kim

Eun

, et al. Bot in the bunch: facilitating group chat discussion by improving efficiency and participation with a chatbot. In: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. New York, NY, USA: Association for Computing Machinery, 2020, pp. 1–13.

38.

Kim

Eun

Seering

, et al. Moderator chatbot for deliberative discussion: effects of discussion structure and discussant facilitation. Proc ACM Hum Comput Interact 2021; 5: 1–26. DOI: 10.1145/3449161.

39.

Lee

Song

, et al. SolutionChat: real-time moderator support for chat-based structured discussion. In: Proceedings of the 2020 CHI conference on human factors in computing systems. New York, NY: ACM, 2020, pp. 1–12.

40.

Liu

Xiao

Browne

, et al. ConsensUs: supporting multi-criteria group decisions by visualizing points of disagreement. Trans Soc Comput 2018; 1(1): 1–26. DOI: 10.1145/3159649.

41.

Dortheimer

Martelaro

Sprecher

Schubert

. Evaluating large-language-model chatbots to engage communities in large-scale design projects. AI EDAM. 2024; 38: e4.

42.

Brown

Mann

Ryder

, et al. Language models are few-shot learners, 2020. https://arxiv.org/abs/2005.14165.2005.14165.

43.

Jones

Petrescu

Till

. Architecture and participation. Oxfordshire, UK: Taylor & Francis, 2005.

44.

Sanoff

. Community participation methods in design and planning. Hoboken, NJ: John Wiley & Sons, 1999.

45.

Francis

. Community design. J Architect Educ 1983; 37(1): 14–19.

46.

Wates

Knevitt

. Community architecture (routledge revivals): how people are creating their own environment. London, UK: Routledge, 2013.

47.

Hamdi

. Housing without houses: participation, flexibility, enablement. London, UK: Intermediate Technology Publications Ltd (ITP), 1995.

48.

Sanoff

. Multiple views of participatory design. Focus 2011; 8(1): 7.

49.

Simonsen

Robertson

. Routledge international handbook of participatory design. London, UK: Routledge, 2012.

50.

Kleinsmann

Valkenburg

. Barriers and enablers for creating shared understanding in co-design projects. Des Stud 2008; 29(4): 369–386.

51.

Fung

. Varieties of participation in complex governance. Publ Adm Rev 2006; 66: 66–75.

52.

Bodker

Kensing

Simonsen

. Participatory IT design: designing for business and workplace realities. Cambridge, MA: MIT Press, 2004.

53.

Roland

Sanner

Sæbø

, et al. P for platform. architectures of large-scale participatory design. Scand J Inf Syst 2017; 29(2): 1.

54.

Mahajan

Hausladen

Argota Sánchez-Vaquerizo

, et al. Participatory resilience: surviving, recovering and improving together. Sustain Cities Soc 2022; 83: 103942.

55.

Soden

Lallemant

Hamel

, et al. Becoming interdisciplinary: fostering critical engagement with disaster data. Proc ACM Hum Comput Interact 2021; 5(CSCW1): 1–27.

56.

Wilson

. Planning and politics: citizen participation in urban renewal. J Am Inst Plan 1963; 29(4): 242–249.

57.

Irani

. “Design thinking”: defending silicon valley at the apex of global labor hierarchies. Catalyst 2018; 4(1): 1–19.

58.

Davis

Niki

Langley

, et al. Low-contact co-design: considering more flexible spatiotemporal models for the co-design workshop. Strategic Design Research Journal 2021; 14(1).

59.

Racadio

Rose

Kolko

. Research at the margin: participatory design and community based participatory research. In: Proceedings of the 13th Participatory Design Conference: Short Papers, Industry Cases, Workshop Descriptions, Doctoral Consortium Papers, and Keynote Abstracts - Volume 2. PDC ’14. New York, NY: Association for Computing Machinery, 2014, pp. 49–52. DOI: 10.1145/2662155.2662188.

60.

Beall

Crankshaw

Parnell

. Uniting a divided city: governance and social exclusion in Johannesburg. London, UK: Routledge, 2014.

61.

Pedell

Vetere

Miller

, et al. Tools for participation: intergenerational technology design for the home. Int J Des 2014; 8(2).

62.

Björgvinsson

Ehn

Hillgren

. Participatory design and “democratizing innovation”. In: Proceedings of the 11th Biennial participatory design conference. New York, NY: ACM, 2010, pp. 41–50.

63.

Marres

. Material participation: technology, the environment and everyday publics. Berlin, Germany: Springer, 2016.

64.

Luck

. What is it that makes participation in design participatory design? Des Stud 2018; 59: 1–8.

65.

McDonnell

. Collaborative negotiation in design: a study of design conversations between architect and building users. CoDesign 2009; 5(1): 35–50.

66.

Dortheimer

. Collective intelligence in design crowdsourcing. Mathematics 2022; 10(4): 539.

67.

Dortheimer

Neuman

Milo

. A novel crowdsourcing-based approach for collaborative architectural design. In: Proceedings of the 38th eCAADe Conference; Education and Research in Computer Aided Architectural Design in Europe, Berlin, Germany, September 2020.

68.

Dortheimer

Yang

, et al. Conceptual architectural design at scale: a case study of community participation using crowdsourcing. Buildings 2023; 13(1): 222.

69.

DiSalvo

Clement

Pipek

. Participatory design for, with, and by communities. Routledge international handbook of participatory design. London, UK: Routledge, 2012, pp. 182–209.

70.

Ahmed

Kok

Howard

, et al. Online community-based design of free and open source software for transgender voice training. Proc ACM Hum Comput Interact 2021; 4(CSCW3): 1–27.

71.

Beniwal

Kachru

Chopra

, et al. What is participatory design? expanding worlds of participatory design from India. In: Proceedings of the participatory design conference 2022, volume 2. New York, NY: ACM, pp. 269–270.

72.

Neumann

Star

. Making infrastructure: the dream of a common language. In: Participatory Design Conference, Cambridge, Ma, January 1996.

73.

Shapiro

. Participatory design: the will to succeed. In: Proceedings of the 4th Decennial Conference on Critical Computing: Between Sense and Sensibility, New York, NY, 20–24 August 2005.

74.

Balka

. Inside the belly of the beast: the challenges and successes of a reformist participatory agenda. In Proceedings of the Ninth Conference on Participatory Design: Expanding Boundaries in Design-Volume 1, New York, NY, 1–5 August 2006.

75.

Seaver

. Care and scale: decorrelative ethics in algorithmic recommendation. Cult Anthropol 2021; 36(3): 509–537.

76.

Lempert

Summerson Carr

. Scale: discourse and dimensions of social life. Berkeley, CA: University of California Press, 2016.

77.

Tsing

. On nonscalability. Common Knowl 2012; 18(3): 505–524.

78.

Bødker

Kyng

. Participatory design that matters—facing the big issues. ACM Trans Comput Hum Interact 2018; 25(1): 1–31.

79.

Simonsen

Hertzum

. Sustained participatory design: extending the iterative approach. Des Issues 2012; 28(3): 10–21.

80.

Iversen

Dindler

. Sustaining participatory design initiatives. CoDesign 2014; 10(3-4): 153–170.

81.

Hertzum

Simonsen

. Effects-driven IT development: a strategy for sustained participatory design and implementation. In: Proceedings of the 11th Biennial Conference on Participatory Design: Participation–the Challenge, 2010. New York, NY: Association for Computing Machinery, pp. 61–70.

82.

Robertson

Simonsen

. Challenges and opportunities in contemporary participatory design. Des Issues 2012; 28(3): 3–9.

83.

Dalsgaard

. Participatory design in large-scale public projects: challenges and opportunities. Des Issues 2012; 28(3): 34–47.

84.

Frauenberger

Foth

Fitzpatrick

. On scale, dialectics, and affect: pathways for proliferating participatory design. In: Proceedings of the 15th Participatory Design Conference: Full Papers-Volume 1, New York, NY, 20–24 August 2018.

85.

Ehn

. Participation in design things. In: Participatory design conference (PDC), Bloomington, IN: ACM Digital Library, 2008, pp. 92–101.

86.

Björgvinsson

Ehn

Hillgren

. Agonistic participatory design: working with marginalised social movements. CoDesign 2012; 8(2-3): 127–144.

87.

Karasti

Infrastructuring in participatory design. In: Proceedings of the 13th Participatory Design Conference: Research Papers-Volume 1, New York, NY, 6–10 October 2014.

88.

Dantec

CAL

DiSalvo

. Infrastructuring and the formation of publics in participatory design. Soc Stud Sci 2013; 43(2): 241–264.

89.

Bjögvinsson

Ehn

Hillgren

. Design things and design thinking: contemporary participatory design challenges. Des Issues 2012; 28(3): 101–116.

90.

Le Dantec

. Infrastructures of digital civics: transportation, advocacy, and mobile computing. In: Infrastructuring publics. Berlin, Germany: Springer, 2019, pp. 169–184.

91.

Dunn

. Participatory gis—a people’s gis? Prog Hum Geogr 2007; 31(5): 616–637.

92.

Soma

Dijkshoorn-Dekker

Polman

. Stakeholder contributions through transitions towards urban sustainability. Sustain Cities Soc 2018; 37: 438–450.

93.

Newton

Glackin

. Using geo-spatial technologies as stakeholder engagement tools in urban planning and development. Built Environ 2013; 39(4): 473–501.

94.

Balestrini

Rogers

Hassan

, et al. A city in common: a framework to orchestrate large-scale citizen engagement around urban issues. In: Proceedings of the 2017 CHI conference on human factors in computing systems, New York, NY: Association for Computing Machinery, 2017, pp. 2282–2294.

95.

Mosconi

Korn

Reuter

, et al. From facebook to the neighbourhood: infrastructuring of hybrid community engagement. Comput Support Coop Work 2017; 26: 959–1003.

96.

Ietto

Rabe

Muth

, et al. Blockchain for citizens’ participation in urban planning: the case of the city of berlin. a value sensitive design approach. Cities 2023; 140: 104382.

97.

Poplin

. Playful public participation in urban planning: a case study for online serious games. Comput Environ Urban Syst 2012; 36(3): 195–206.

98.

Schouten

Ferri

de Lange

, et al. Games as strong concepts for city-making. In: Playable cities: The city as a digital playground. Berlin, Germany: Springer, 2017, pp. 23–45.

99.

Ślusarczyk

Strug

Paszyńska

, et al. Semantic-driven graph transformations in floor plan design. Comput Aided Des 2023; 158: 103480.

100.

Jiang

Webster

, et al. Building layout generation using site-embedded gan model. Autom ConStruct 2023; 151: 104888.

101.

Stouffs

Biljecki

. Generative adversarial networks in the built environment: a comprehensive review of the application of gans across data types and scales. Build Environ 2022; 223: 109477.

102.

Liao

Huang

, et al. Automated structural design of shear wall residential buildings using generative adversarial networks. Autom ConStruct 2021; 132: 103931.

103.

Castro Pena

Carballal

Rodríguez-Fernández

, et al. Artificial intelligence applied to conceptual design. a review of its use in architecture. Autom ConStruct 2021; 124: 103550.

104.

Chaillou

. Archigan: artificial intelligence x architecture. In: Architectural intelligence: selected papers from the 1st international conference on computational design and robotic fabrication (CDRF 2019). Berlin, Germany: Springer, 2020, pp. 117–127.

105.

Parente

Rodrigues

Rangel

, et al. Integration of convolutional and adversarial networks into building design: a review. J Build Eng 2023; 76: 107155.

106.

Pal

Basu

. Artificial intelligence in architecture: generating conceptual design via deep learning. Int J Architect Comput 2018; 16(4): 306–327.

107.

Tamke

Nicholas

Zwierzycki

. Machine learning for architectural design: practices and infrastructure. Int J Architect Comput 2018; 16(2): 123–143.

108.

McDonnell

Lloyd

. About designing: analysing design meetings. Boca Raton, FL: CRC Press, 2022.

109.

Luck

. Dialogue in participatory design. Des Stud 2003; 24(6): 523–535.

110.

Luck

McDonnell

. Architect and user interaction: the spoken representation of form and functional meaning in early design conversations. Des Stud 2006; 27(2): 141–166.

111.

Luck

. Learning to talk to users in participatory design situations. Des Stud 2007; 28(3): 217–242.

112.

Oak

. What can talk tell us about design? analyzing conversation to understand practice. Des Stud 2011; 32(3): 211–234.

113.

Habermas

. The structural transformation and the public sphere: an inquiry into a category of bourgeois society. Cambridge, MA: MIT Press, 1989.

114.

Rawls

. Political liberalism. Cambridge, UK: Cambridge University Process, 1996.

115.

DiSalvo

. Adversarial design. Cambridge, MA: MIT Press, 2015.

116.

Mouffe

. Deliberative democracy or agonistic pluralism? Soc Res 1999; 66: 745–758.

117.

Mouffe

. Agonistics: thinking the world politically. London, UK: Verso Books, 2013.

118.

Androutsopoulou

Karacapilidis

Loukis

, et al. Transforming the communication between citizens and government through ai-guided chatbots. Govern Inf Q 2019; 36(2): 358–367.

119.

Jain

Kumar

Kota

et al. Evaluating and informing the design of chatbots. In: Proceedings of the 2018 ACM conference on designing interactive systems. New York, NY: ACM, 2018, pp. 895–906.

120.

Han

Zhou

Turner

, et al. Designing effective interview chatbots: automatic chatbot profiling and design suggestion generation for chatbot debugging. In: Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems, Yokohama, Japan, 8–13 May 2021.

121.

Dou

Zhu

et al. One chatbot per person: creating personalized chatbots based on implicit user profiles. In: Proceedings of the 44th international ACM SIGIR conference on research and development in information retrieval. New York, NY: ACM, 2021, pp. 555–564.

122.

Putri

Meidia

Gunawan

. Designing intelligent personalized chatbot for hotel services. In: Proceedings of the 2019 2nd International Conference on Algorithms, Computing and Artificial Intelligence, Cambridge, Ma, January 1996.

123.

RIBA . RIBA plan of work 2020 overview, 2020. https://www.ribaplanofwork.com.

124.

Zafer-Smith

. A sheltered place, a sheltered person? ordering wellbeing in a time between wars in the northern galilee. Anthropol Med 2021; 28(1): 28–46.

125.

Rabinowitz

Khawalde

. Demilitarized, then dispossessed: the Kirad Bedouins of the Hula Valley in the context of Syrian–Israeli relations. Int J Middle East Stud 2000; 32(4): 511–530.

126.

Maxwell

. Qualitative research design: an interactive approach. Thousand Oaks, CA: Sage publications, 1996.