The Center on Information and Narrative Complexity

About:

The Center on Information and Narrative Complexity (CINC) works to identify and mitigate the impacts of strategic influence campaigns from foreign adversaries intent on undermining the interests and security of the United States (U.S.) and allies. Influence campaigns are a significant component of global competition today, used by competitor nations and adversarial non-state actors to sow confusion or mold public opinion in their favor or against U.S. interests. The goals can vary — ranging from reducing confidence in the U.S. as an ally to preemptive justification for military actions. Geopolitical influence operations are often inexpensive and low-risk but can cause real damage to a nation’s ability to respond to strategic threats.

CINC applies a complex systems approach to study this phenomenon in today’s rapidly-evolving information environment. With the broad deployment of generative AI tools that could supercharge influence campaigns, the center is helping develop new tools to identify AI-generated content and to leverage large language models in support of U.S. national security.

Capabilities

We develop innovative complex systems solutions that bridge the gap between advanced artificial intelligence and how people process information in the context of their personal and cultural narratives. Our capabilities include:

Detecting and countering AI-driven information manipulation
Integrating interdisciplinary insights to address complex technical challenges in society

Strengthening resilience in civil society against information threats
Enhancing the effectiveness of strategic communications

Featured projects

CONNECT

Aims to combine cutting-edge technology and cultural insights to advance strategic digital communication. By refining the capabilities of Large Language Models for narrative analysis and modulation, we aim to enhance messaging efficacy across diverse cultural landscapes. This project focuses on developing advanced techniques to encode, decode, and modulate narratives tailored to specific audiences, leveraging both computational and theoretical approaches.

NewsWell

This project explores how AI-driven tools can bolster community-based journalism, strengthening the sustainability of local news outlets while ensuring towns and cities across the country have reliable, trusted information for informed civic engagement and a robust democracy.

Minerva

This Department of Defense project illuminated important aspects of China’s influence operations across Southeast Asia. By integrating social listening techniques with on-the-ground expert insights, this project uncovered differences between online narratives and offline realities, enabling more nuanced understanding of local public opinion in Indo-Pacific countries and better-informed strategic responses.

Semantic Information Defender (SemaFor)

This recently-completed project addressed the growing challenge of distinguishing human-created content from AI-generated and manipulated media in the era of generative AI by developing advanced algorithms specifically designed to detect, attribute, and characterize AI-generated media. Despite the rapid advancement of Large Language Models our team successfully delivered state-of-the-art text detection capabilities, achieved top performance metrics across rigorous text analytics evaluations and published numerous papers across top computer science venues.