Recent research · ICWSM 2025
Online myths on opioid use disorder: Reddit vs. large language models
Shravika Mittal, Hayoung Jung, Mai ElSherief, Tanushree Mitra, Munmun De Choudhury
The Center for Responsibility in AI Systems & Experiences advances artificial intelligence that serves the public good, builds trust, and supports open innovation. Our work centers on three aims that define responsible AI in practice.
applied AI
epistemic foundations
open & specialized systems
Three research aims
Responsible AI in practice
IAim I
Applied AI in Underserved Areas
We study applications of AI in areas underserved by market incentives, especially in the public sector and the physical, life, and social sciences. We aim to improve equitable access and use as the world adapts to the proliferation of AI.
IIAim II
Epistemological Foundations of AI
We study how AI systems generate and justify claims across pre-training, mid-training, and post-training stages. Our focus is on the epistemological foundations of trust: how errors arise, when abstention is the right choice, how uncertainty should be communicated, and how explanations can support human oversight. By grounding trust in epistemology, we aim to create systems aligned with human values.
IIIAim III
Open & Specialized Systems
Large proprietary models limit transparency, adaptability, and governance. We advance smaller, open-weight, and open-source models that can be adapted and fine-tuned in specialized settings. These models reduce dependency on closed APIs, lower costs, and enable communities to govern AI systems according to their needs.
Selected work
Recent publications
Bidirectional Human-AI Alignment: Emerging Challenges and Opportunities
CHI 2025 · Extended AbstractsRead ↗
Epistemic Alignment: A Mediating Framework for User–LLM Knowledge Delivery
COLM 2025Read ↗
MythTriage: Scalable Detection of Opioid Use Disorder Myths on a Video-Sharing Platform
EMNLP 2025Read ↗
Mind the Value-Action Gap: Do LLMs Act in Alignment with Their Values?
EMNLP 2025Read ↗
Towards Designing Social Interventions for Online Climate Change Denialism Discussions
CSCW 2025Read ↗
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
ACM FAccT 2021Read ↗
Leadership
Meet the full team ↗
Our co-founding directors
Bill Howe
Co-Founding Director
Associate Professor in the iSchool, Adjunct Associate Professor in Computer Science & Engineering, and Founding Associate Director of the eScience Institute.
Tanu Mitra
Co-Founding Director
Associate Professor at the Information School, where she leads the Social Computing research group studying and building large-scale social computing systems to counter problematic information online.
Chirag Shah
Co-Founding Director
Professor in the Information School and Founding Director of the InfoSeeking Lab, focusing on information seeking, human-computer interaction, and social media.
In partnership with
Get involved
Join the RAISE community
We believe in the power of collaboration. Join us in building AI systems and experiences with fairness, accountability, transparency, and ethics — students, researchers, and partners are all welcome.