Social media feeds and AI systems now sit between people and much of what they read, learn, and decide. My research asks: how do these systems shape people's wellbeing, trust, and behavior, and how can we design technology that people use more intentionally?
My work sits at the intersection of computational social science, HCI, and NLP.
I use causal inference on large-scale social media data to understand how what people see and engage with online changes their mental health, their language, and the communities they form.
I study when and why people trust AI, and whether that trust matches how much they actually rely on it, drawing on interviews, think-aloud studies, and audits of how LLMs respond in sensitive settings.
At IIT Kharagpur I worked on EEG-based classification systems that perform well with very little training data.
- Study how algorithmic curation on Bluesky shapes what people see, and how exposure to curated feeds changes the way they write over time.
- Examine the causal effects of engaging with news on Bluesky on users' psychosocial wellbeing, using propensity score matching.
- Conduct semi-structured interviews and think-aloud studies on how students trust and rely on AI, drawing on the Mutual Theory of Mind framework.
- Contribute to audits of LLM caregiving support and to studies of AI agent communities.
- Worked with Prof. Juhi Kulshrestha (May to August 2026) on web tracking panel data.
- Studied which properties of news content, such as framing, topic, and political congruence, drive its effects on mental health.
- Explored causal relationships between late-night browsing, sleep quality, and other mental health factors.
Research: Currently working on AAINA, a usability study of a prototype that helps people reflect on their emotional wellbeing while using social media.
Always happy to chat about research, collaborations, or anything at the intersection of technology and wellbeing. Feel free to reach out!
opal2@illinois.edu