Advancing Human-Centered AI for Behavioral Health

Explore cutting-edge research tools and interactive demos developed by the Chat Lab to advance human-centered AI for behavioral health and social impact. Our work focuses on bridging the gap between automated intelligence and expert human judgment, creating systems that are both accurate and interpretable for real-world applications.

On this page, you’ll find video demonstrations showcasing innovative frameworks designed to tackle complex challenges such as mental health signal detection in social media and behavioral health identification from emergency response narratives. These tools integrate domain expertise, advanced machine learning, and human-AI collaboration to deliver scalable, reliable, and transparent solutions for sensitive contexts.

 

Domain-Driven AI for Behavioral Health Analysis in Emergency Reports


This video demonstrates an automated behavioral health identification framework designed specifically for the linguistic complexity of first responder reports. Traditional natural language processing and deep learning approaches struggle in this domain due to domain-specific language and overlapping conditions.

 

 

Human-AI Collaboration for Detecting Mental Health Signals in Social Media


This video presents a human-AI collaborative framework for analyzing mental health–related signals in social media data. Due to the linguistic diversity and contextual complexity of platforms like Reddit, conventional machine learning models often struggle to capture meaningful psychological cues.