Identity, Safety, and Platform Governance

Examining how digital platforms shape identity, expression, and safety for historically marginalized communities across South Asia, through decolonial and sociotechnical perspectives.

← All research projects

Research Summary

This project examines how digital platforms shape identity, expression, and safety for historically marginalized communities across South Asia and its diaspora. I study how platform governance, algorithmic systems, and colonial histories shape whose identities become legible, whose voices are amplified, and who is made unsafe, from Bengali communities reclaiming narrative agency online to religious and Indigenous minorities facing hate, fear, and surveillance.


Why this research?

Recommendation algorithms, moderation systems, interface designs, and economic incentives influence whose voices are heard online. Much of the work on online communities in HCI and social computing comes from North American and European contexts, so how colonial histories, linguistic diversity, religious identities, and regional politics shape participation elsewhere remains insufficiently understood. Minority users, in particular, often experience harms that platform policies were not designed to recognize: implicit erasure, misrepresentation, and fear rooted in local histories of violence, alongside surveillance exercised by peers and kin as well as states.

Rather than treating identity as a fixed characteristic, I examine it as an ongoing sociotechnical accomplishment that emerges through interactions among people, platforms, algorithms, and historical power structures.

Questions guiding this work include:

  • How do language, history, and colonial legacies shape identity and participation online?
  • How do creators and communities navigate platform governance, moderation, and algorithmic visibility?
  • How do fear, surveillance, and social conformity silence minority voices online?
  • What would platform governance grounded in minority perspectives and restorative justice look like?

Research Approach

  • Interviews and digital ethnography with community members, content creators, and activists
  • The asynchronous remote community method to engage distributed minority populations
  • Computational, qualitative, and critical discourse analyses
  • Decolonial, postcolonial, and sociotechnical perspectives

Research Outcomes and Contributions

Identity and platform governance

  • Identity decolonization online: Bengali communities on Quora collaboratively reclaim narrative agency over their identities, and platform governance shapes which linguistic and cultural identities a platform comes to hold. (See CHI 22, CSCW 21, and my PhD dissertation)

  • Decolonial discourse on YouTube: Creators engaging in Bengali decolonial discourse across Bangladesh, India, and Pakistan use video to reimagine nationalism and community, while navigating audience backlash, differential monetization, opaque moderation, copyright regimes, and state influence routed through intermediaries. (See CSCW 24, GROUP 27)

  • Creators and identity work: Facebook microcelebrities engage strategically in sociopolitical discourse, and tutors on transnational gig-education platforms align their identities to client expectations through new forms of emotional labor. (See ICTD 22, JCSCW 25)

Safety for minority and Indigenous communities

  • The politics of fear: Bangladeshi religious minorities’ use of social media is shaped by fear tied to social conformity, misinformation, stigma, and South Asian postcolonial memory, pushing minority voices into a “spiral of silence.” (See CSCW 24)

  • Usership: Drawing an analogy to citizenship, we analyze the status, rights, membership, and participation of intersectionally marginalized users, such as Indigenous ethnic minorities and women, facing hate speech online. (See ICTD 24)

  • Lateral surveillance and resilience: During Bangladesh’s 2024 uprising, protesters experienced surveillance from peers, kin, and intimate ties as well as the state, and responded with improvised encryption, panic deletions, satire, and hidden speech. (See USENIX Security 26)

  • Restorative moderation: We propose moving from punishment-based moderation toward restorative justice that addresses hermeneutical injustice, demonstrated through Mod-Guide, an LLM-based moderation feedback system grounded in Hindu and Chakma community knowledge. (See COMPASS 26)

Global and decolonial HCI

My work also helps build communities of scholarship on decolonial computing and HCI across borders, and questions how Western heteropatriarchal writing norms shape scientific communication in HCI. (See CHI 25, CSCW 22 Poster, CSCW 23 Workshop)


Current Directions

  • AI-mediated identity construction and cultural representation
  • Pluralistic moderation for multilingual and multi-faith platforms
  • Safety, anonymity, and collective resilience in repressive contexts
  • Digital public spheres in the Global South