NSF awards RIT more than $1 million to develop digital media forensics platform

RIT researchers aim to tackle misinformation by detecting digital deception and deepfakes

Scott Hamilton/RIT

A team led by Endowed Professor of Cybersecurity Matthew Wright, center, received an NSF grant to help combat misinformation and manipulated media, including deepfakes. The project in development, called Varuna, will organize digital media forensics tools for journalists and analysts.

A new RIT project aims to bring order to the chaotic world of digital media forensics.

Researchers were awarded a total of $1.5 million from the National Science Foundation for “VARUNA: A User-Centered All-in-One Media Forensics Platform.” RIT was awarded $1,125,000 for the Varuna project, with partner researchers at University of Mississippi and Michigan State University earning additional funding.

face made of multimedia

To learn more

Read more about how RIT experts are helping journalists and intelligence analysts detect digital deception.

Distinguishing authentic content from altered content has become increasingly difficult as digital manipulation technologies—including AI-generated deepfakes—become more sophisticated. To signify their dedication to structure, the researchers named their Varuna project after the Hindu god of cosmic order.

The project is dedicated to designing an all-in-one digital media forensics platform that meets the specific needs of journalists and forensic analysts. Varuna will organize analysis tools and models into a single, streamlined system that enhances efficiency and ease of use.

“Central to Varuna’s development is our commitment to working closely with users so we can empower media and forensics professionals to effectively tackle misinformation,” said Matthew Wright, principal investigator of the project and endowed professor and chair of RIT’s Department of Cybersecurity. “By actively engaging with journalists and analysts, we aim to understand and address their unique challenges and requirements.”

In Varuna, users will import images, audio, or video files. Preliminary analysis will help users identify which analytics could be most applicable. The user could then dig deeper based on their own experience and the properties of the analytics, such as effectiveness for specific types of media or the time it takes to run.

The tool will provide explanations about what to expect from analytics. The tool will generate reports with numbers and text descriptions of the findings that help users understand the results. Experts can use the tool’s ontology to narrow down why content needs to be analyzed, where in the media analysts should focus their attention, and what artifacts they should look for.

The NSF funding is part of Phase 2 of the Responsible Design, Development, and Deployment of Technologies (ReDDDoT) program. ReDDDoT aims to ensure that ethical, legal, and societal considerations and community values are embedded across technology lifecycles to generate products that promote the public’s wellbeing and mitigate harm. ReDDDoT projects all focus on key technologies specified in the CHIPS and Science Act.

RIT experts working on Varuna, include Wright, and co-PIs Christopher Schwartz, research scientist in the Department of Cybersecurity, David Schwartz, director of the School of Interactive Games and Media, and Ersin Uzun, endowed executive director of the ESL Global Cybersecurity Institute.

Varuna is an extension of the DeFake Project, a team of RIT student and faculty researchers leading the charge to detect deepfakes. The DeFake Project has more than $2 million in funding from the NSF and Knight Foundation.


Recommended News