Broadly, I explore questions of human subjectivities in human-technology partnerships. To this end, I research how people (1) make sense of technological changes and (2) participate in the future-making of technologies. Currently, I have my eye on speed and scale as they relate to cultures of AI:
(1) How do cultural workers make sense of the speeding up of content creation?
From smart design templates, automated audio mastering, to new generative AI tools, emerging technologies set to speed up content creation. Amidst such shifts, I study how cultural workers navigate the desired role of automation, authenticity, and the sociotechnical assemblage for such configuration. To this end, my research projects span the realms of social media content creation, music production, and filmmaking.
(2) How can people meaningfully participate in shaping our futures at the scale of AI?
Moving beyond the consensus that people should participate in shaping our futures with AI, I explore if and how careful human intervention is possible at the scale of AI. Beyond participation in the form of design workshops, data annotation, or model training, I conceptualize and operationalize new approaches to participatory AI across timescales and beyond project lifecycles.
Before my PhD, I explored how AI and digital platforms reshape notions of authenticity and expertise across a variety of socio-technical contexts, from participatory AI design, architectural design, clinical decision-making, underground music scenes, to networked activism.
Mapping the Participatory Turn in AI Design
March 2021 - September 2023
Collaborator: Fernando Delgado, Michael Madaio, Qian Yang
Despite the growing consensus that stakeholders affected by AI systems should participate in their design, enormous variation and implicit disagreements exist among current approaches. For researchers and practitioners who are interested in taking a participatory approach to AI design and development, it remains challenging to assess the extent to which any participatory approach grants substantive agency to stakeholders.
Stage 1 - Conceptual Framework: We first derive a conceptual framework through synthesis of literature across technology design, political theory, and the social sciences that researchers and practitioners can leverage to evaluate approaches to participation in AI design.
Stage 2 - Empirical Foundation: We then evaluated the current state of stakeholder participation in AI. We developed a corpus including 80 research publications that described AI projects in which stakeholder participation was pursued, as well as 12 interviews with researchers and industry practitioners who were authors on these papers.
Fernando Delgado*, Stephen Yang*
, Michael Madaio†
, and Qian Yang†
. (2023). The Participatory Turn in AI Design: Theoretical Foundations and the Current State of Practice. Paper accepted to the ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO'23). [link
Fernando Delgado, Stephen Yang
, Michael Madaio, and Qian Yang. (2021). Stakeholder Participation in AI: Beyond "Add Diverse Stakeholders and Stir." Paper Presented at the Human-Centered AI Workshop at NeurIPS 2021. [link
* Co-first authors and co-last authors contributed equally.
When Design Workshops Meet Chatbot: Operationalizing Participatory Architectural Design
January 2022 - August 2022
Collaborator: Jonathan Dortheimer, Aaron Sprecher, Qian Yang
Today’s participatory design (PD) processes focus on face-to-face design workshops as the primary method of participation. However, their time- and resource-intensive nature poses practical constraints on the extent to which these workshops can amplify diverse voices. With the maturing capabilities of large language models (LMs) like GPT-3, chatbots present great promise as an emerging PD method that can distribute participation across geographic, resource, and language differences. To this end, we facilitated participatory architectural design project in northern Israel through a hybrid PD workflow that leveraged both (1) group-based, face-to-face design workshops and (2) one-on-one, remote, GPT-3-powered chatbot engagement. Our experience offers a valuable reference for future research that seeks to incorporate chatbots in PD processes or explore spatial-temporal possibilities of PD beyond traditional face-to-face workshops.
, Jonathan Dortheimer, Qian Yang, Aaron Sprecher. When Design Workshops Meet Chatbots: Exploring a Hybrid Approach to Community Participation. Under Review.
Jonathan Dortheimer, Stephen Yang,
Qian Yang, Aaron Sprecher. (2023). Conceptual Architectural Design at Scale: A Case Study of Community Participation Using Crowdsourcing. Buildings
, 222. [link
Harnessing Biomedical Literature to Calibrate Clinicians' Trust in AI-Powered Decision Support Systems
January 2021 - January 2023
Collaborator: Yiran Zhao, Qian Yang, Yuexing Hao, Kexin Quan, Fei Wang, Volodymyr Kuleshov, Bojian Hou
Clinical decision support tools (DSTs), powered by Artificial Intelligence (AI), promise to improve clinicians' diagnosis and treatment decision-making. However, no AI model is always correct. DSTs must enable clinicians to validate each AI suggestion, convincing them to take correct suggestions while rejecting errors. Towards this goal, existing DST designs often explain AI's inner workings or performance indicators. We chose a different approach: We investigated how clinicians validated each other's suggestions in practice.
As a first step, we conducted 12 interview with clinicians to investigate how clinicians sought, prioritized, and synthesized information from the literature to validate care suggestions. We paid particular attention to whether and how clinicians' behaviors deviated from best practices in the time pressure of patient care.
Building on our insights, we designed a new DST that embraces clinicians' naturalistic evidence-calibration behaviors. The web-based prototype
can be accessed here. This design helps clinicians deliberate whether to take an AI's suggestion by providing a list of supporting and opposing evidence from biomedical literature.
Qian Yang, Yuexing Hao*
, Kexin Quan*
, Stephen Yang*
, Yiran Zhao*
, Volodymyr Kuleshov, and Fei Wang. Harnessing Biomedical Literature to Calibrate Clinicians’ Trust in AI Decision Support Systems. In Proceedings of the 2023 ACM Conference on Human Factors in Computing Systems (CHI’23). [link
* Co-second authors contributed equally listed in alphabetical order.
Playing with Visibility in Underground Electronic/Dance Music Scenes
March 2021 - January 2023
The mediation of mobile and social media has reshaped how people imagine, understand, and, in turn, negotiate the visibility of their self-expression, information sharing, and relationship-building. Prior work on visibility management has focused on the online circulation of content without attending to the spatial-temporal context of situated technology use. To investigate these dimensions of visibility management on and through mobile and social media, I examined how participants in underground electronic/dance music culture (EDMC) manage their visibility. This research examines how the mediation of mobile and social media has reshaped what participants of underground EDMC do to maintain their shared culture of secrecy.
I conducted 20 nights of field observations at live music events and 27 semi-structured interviews with scene participants. I adopted a multi-site approach by studying four underground dance music scenes in distinct socio-political contexts––New York City; Ithaca, New York; Taipei, Taiwan; Berlin, Germany. Taking a “field site as network” approach (Burrell, 2009), I selected the four music scenes as “entry points” to gain access to underground EDMC.
Stephen Yang. (Accepted and forthcoming). Playing with Visibility: Underground Electronic/Dance Music in the Smartphone Age. International Journal of Communication.
Stephen Yang. (Accepted and forthcoming). The Platformized Rhythms of Underground Electronic/Dance Music: Re-Configuring the Temporal Experience of the Nightscape In P. Ballon & Smets, A. (Eds). To Appear in Handbook of Platform Urbanism. Edward Elgar.
#IChooseFish and the New Frontier of Civic Participation in Vietnam
Jan 2021 - May 2022
Collaborator: Emily La
In April, 2016, millions of dead, toxic fish washed up along the north-central Vietnamese coast. In the aftermath, the Vietnamese official gave a statement: “Choose fish or steel” to imply that Vietnam’s economic growth is not without some sacrifice to the environment. In response, the hashtag #IChooseFish became viral, through which the Vietnamese public shared information on upcoming protests and coordinated emergency relief for local residents. This project seeks to to understand the role of technologies for dissents in non-Western, authoritative contexts.
We adopted a mixed-method approach to underpin how the use of hashtags and mobile media contributed to the discourse formation surrounding the #IChooseFish movement. On the quantitative side, we drew on computational analysis of Twitter data to examine hashtag co-occurrence patterns and the specific discourse that have developed from this practice. On the qualitative side, we drew on semi-structured interviews with protesters to attend to how they leverage hashtags and mobile social media for their participation in the protests.
Emily La, Stephen Yang (May, 2022). The Role of Hashtags and Mobile Media in Shaping Discourse During the #IChooseFish Environmental Movement. Presented 72nd annual meeting of the International Communication Association (ICA). Paris, France.
Documenting the Lived Experiences of Data Subjects
January 2020 - January 2021
Collaborator: Ranjit Singh, Amy Eng, Deana Gonzales, Ciarra Lee, Emma Li, Cassidy McGovern, Annika Pinch, Sterling Williams-Ceci, Kyra Wisniewski,, and Malte Ziewitz
Being ranked in web search engines can make or break or business, reputation, or career. Big companies and organizations may have the resources to hire PR and marketing consultants to affect their standing in the engines. But how do less well-resourced people make sense and cope with their predicament? Through a series of qualitative interviews, we collected the stories from small business owners and local activists to politicians running for public office and people re-entering society after prison terms.
R. Singh, A., Eng, D. Gonzales, C. Lee, E. Li, C. McGovern, A. Pinch, S. Williams-Ceci, K., Wisniewski, S. Yang, and M. Ziewitz. (2020). Scoring Struggles: Everyday Experiences of Web Search Engines. Working Paper #1,
August. Ithaca, NY: Digital Due Process Clinic.
Flo: Designing Mood-Aware Space for Co-Listening
January 2021 - May 2021
Collaborator: Tony Bridges, Emma Wang, Laurel Prime
Existing solutions for music co-listening, the act of listening to music together synchronously, has neglected the particular social and emotional needs of people with anxiety and/or depressive disorders. To bridge the gap in the design space, this project seeks to address such needs with a design for co-listening. Our research and design process attempts to valorize the thoughts, experiences, and feelings of this underserved population.
Through semi-structured interviews, diary study, and a co-design workshop, we found that the moods of our users are strongly influenced by the music they listen to, and often feel a lack of control over the emotional dimension of their co-listening experience. This makes it challenging for them to pursue social connectedness through music co-listening. Based on such insights, we designed Flo, a music-oriented virtual space that offers mood-based curations for co-listening to foster music-led sociality through emotional sensibility.
We have developed a high-fidelity prototype
of Flo on Figma and a written report
that walks through the design process.
Lively: Performance Streaming for Amateur Musicians
September 2020 - December 2020
Collaborator: Donna Ilyukh, Danni Ji, Kepler Despinos
This project investigates the needs of the needs of amateur, independent, up-and-coming musicians that are unable to perform at physical venues during the pandemic. Through semi-structured interviews, card sorting , and affinity diagramming, we developed Lively, a streaming platform that allows such musicians to stream their performances on their own. Our design highlights two key dimensions of their needs - (1) the needs for hassle-free control over the technical aspects of performance and (2) the needs to engage with their audience throughout the performance.
We have developed a high-fidelity prototype
of Lively on Figma and a written report
that walks through the design process.