We bring together humanists and technologists to tackle some of the most complex communication challenges facing the media, education, public health, and government. Projects at the Center range from exploratory research in the lab to field pilots of advanced prototypes with potential for scale.
Some of our work within the research lab and the prototyping workshop is highlighted below.
A story-telling card game generated from community data which helps make speaking personally and emotionally easy and fun
A 2-person digital experience which supports intimate conversation by visualizing shared values using a text-to-image model.
Narrative Reframing Agents
Design of narrative agents to assist in reframing traumatic experiences.
Real Talk Living Lab in Boston
A new civic infrastructure in Boston grounded in dialogue as a way of building up our “civic muscles” of democracy
A community-based civic space for nuanced and authentic dialogue among the MIT community
CCC High School Internship Program
Empowering student leaders to design and facilitate constructive conversation spaces
A tech-enhanced Citizens’ Assembly pop-up lab
Tech-enhanced Citizens’ Assemblies toward a more inclusive, healthy, and constructive democracy
Project Odessa: Decentralized social systems in theory and implementation
Project Odessa builds on theoretical designs and provides a mobile app sandbox for such experimental implementations.
Producing speech that de-identifies the speaker’s voice characteristics while preserving naturalness, intelligibility, and paralinguistic features like emotion + prosody.
Exploring the ways in which we respond differently to statements of opinion and statements of experience.
Assessing Affective Engagement with Narratives of Invisible Disability
Deconstructing and reducing situational stigma on the basis of visual and semantic assumption mechanisms.
Narrativity Rating Scale
Development of annotation-based scale to identify, isolate, and measure the emotional-behavioral effects on reader-listeners based on narrative-form elements within written or spoken texts of any type.
City meeting transcript analysis
Can discussions at city council meetings (the LocalView corpus) be used to explain/predict social measures of the cities where they occur? Can it be useful in other CCC projects?
Exploring LLM capabilities to explain concepts/definitions to various audiences
LLMs to characterize groups
Using LLMs (originally trained / fine tuned, now prompted) to answer questions about group beliefs.
Chinese Restaurant Names
Using Language Models to Investigate Cultural Adaptation of Chinese Immigrants from Parallel Chinese Restaurant Names in the United States
AI Sensemaking (Prototyping)
Exploring how LLMs can be utilized to build AI power tools that can assist in a human-led AI-assisted sensemaking process
Discover conversation themes via clustering, relate those to human generated highlights, and explore temporal patterns in how conversations develop.
Exploring conversation health and dynamics using natural language processing and data visualization
Visualizing conversations based on themes
Understanding LLM cultural and demographic knowledge
Probing LLMs to understand gaps in their knowledge around certain traits (gender, race, etc.)
Data Provenance for AI
A massive audit of 1800+ AI text datasets, tracing their provenance and composition from origin to creation.
Characterizing affective polarization on Twitter and Reddit – how it has evolved over time and which topics incite more polarized responses.
Developing a new joint model combining graph + text representations. Evaluated on social media (Twitter), citation networks (Pubmed) and Wikipedia.
Social Media vs Radio
Exploring how social media and radio react to news events
Beat The Virus
How could we deliver trusted public health information to millions of people in engaging ways in the pandemic’s very noisy early days? Our answer – #BeatTheVirus (…