Participatory AI

Participatory approaches aim to bring more perspectives—especially those that are historically not represented—into the design and development of technology. Participatory approaches hold promise to shift harmful power imbalances and enable more grassroots and community-driven technology practices. But what does meaningful participatory AI actually look like in practice? How do we avoid risks of "participation-washing"? What tensions arise between considerations of context-specificity versus those of scale? Our work in this area spans case studies and participatory parternships grounded in context, as well as broader frameworks of what participation could and should look like in today's AI landscape.

Counterdata, resistance, agency

What are the ways in which individuals and communities assert agency in response to highly-centralized and black-box algorithmic systems? We are interested in understanding and supporting everyday practices of resistance, subversion, contestation, and counterdata collection. For example, as part of the cross-institutional Counterdata Network, we contribute to understanding the data practices of activists who collect counterdata, and co-design AI-based tools to support their work monitoring human rights abuses.

Sociotechnical evaluation

Evaluation paradigms influence which research agendas are pursued, what is considered "state-of-the-art," and how we prioritize values in technology. How can evaluation methods stem from the lived experience and expertise of those embedded in a downstream context? What does a sociotechnical approach to evaluation—one that is not only focused on model behavior, but on the wider social, political and organizational dynamics at hand—look like? Our work in this area includes problematizing default notions (e.g., of what it means to be an "expert") and designing new systems and practices for contextual, human-centered evaluation.

Participation in the age of foundation models.
ACM Conference on Fairness, Accountability, and Transparency (FAccT ‘24).
Harini Suresh*, Emily Tseng*, Meg Young*, Mary Gray, Emma Pierson, Karen Levy.
[paper]

Co-liberation through Data & Sociotechnical Systems? Reflections, Tensions, and Possibilities.
A Toolbox of Feminist Wonder Workshop, ACM Conference on Computer-Supported Cooperative Work and Social Computing (CSCW ’23).
Harini Suresh*, Nikki Stevens*, Amelia Lee Doğan.
[paper]

Kaleidoscope: Semantically-grounded, context-specific ML model evaluation.
CHI Conference on Human Factors in Computing Systems (CHI ’23).
Harini Suresh, Divya Shanmugam, Annie Bryan, Tiffany Chen, Alexander D’Amour, John V. Guttag, Arvind Satyanarayan.
[paper]

Towards Intersectional Feminist and Participatory ML: A Case Study in Supporting Feminicide Counterdata Collection.
ACM Conference on Fairness, Accountability, and Transparency (FAccT ‘22).
Harini Suresh, Rajiv Movva, Amelia Lee Dogan, Rahul Bhargava, Isadora Cruxên, Ángeles Martinez Cuba, Giulia Taurino, Wonyoung So, and Catherine D’Ignazio.
[paper]

If you are currently an undergraduate or graduate student at Brown interested in working with the lab, reach out directly to Harini at harini_suresh@brown.edu, including some background on yourself and your interests. Please also include if you have a specific thesis or research project you'd like to propose. Our research is interdisciplinary and touches many subject areas -- please feel free to reach out if you think there's a good fit, even if you are outside of the CS department/concentration.

If you are a prospective PhD student, please apply to Brown's CS PhD program directly. You can also email harini_suresh@brown.edu, including some background on why you are interested in the lab. Due to volume, Harini may not be able to reply to all emails about PhD admissions, but all applications will be reviewed carefully.

If you are an individual or organization (in or outside of academia) interested in partnering on a project, please reach out. We are always interested in ways we can support local/grassroots initiatives that align with our mission.