We use natural language processing and computational social science to understand lived experiences and design language technologies that are socially aware and responsive to real-world needs.
We are always interested in exploring innovative research directions at the intersection of language, AI, social behavior, and consequential real-world domains. If you see a promising connection with your work or have an idea that could benefit from interdisciplinary collaboration, please reach out.
We develop computational methods to understand how people describe substance use, recovery, stigma, sensitive experiences, and support across digital spaces, and how language technologies can reduce rather than reproduce harm.
We study AI systems that participate in emotionally sensitive interactions, connecting model behavior to therapeutic processes, human judgments of empathy, and evidence from how people actually use AI for mental-health support.
We examine how generative AI changes judgments of credibility, competence, authority, and truth, and how people negotiate trust, delegation, deception, and belief in AI-mediated environments.
We model consequential online discourse beyond simple binary labels, capturing stance, risk, morality, emotion, user intent, uncertainty, and information behavior across social and health domains.
Alongside the four core pillars, we develop methods that connect NLP with computational social science, human-centered evaluation, scientific communication, and domain knowledge. This work creates methodological bridges that can travel across research areas rather than being tied to a single application domain.