Detect Fakes

How do you spot a deepfake? How can you tell if an image is AI-generated? How well can people distinguish between AI-generated media and images, audio, and video recorded by a traditional camera or microphone? Detect Fakes is research project hosted at Northwestern University by researchers at the Kellogg School of Management to examine how people distinguish truth from fiction in online media, especially as synthetic media becomes more and more realistic. This research started at the MIT Media Lab in 2019 by examining comparing human, AI, and human + AI performance at detecting deepfakes. The first peer-reviewed paper from the Detect Fakes research program was published in 2022 in the Proceedings of National Academy of Science (PNAS). Stay tuned for more research.

The goal of this research program is to reveal what synthetic media looks like, benchmark how well people can distinguish synthetic media from "real" media across a variety of contexts, and generate insights for how to help people distinguish between the two.

Research Team

Negar Kamali, Angelos Chatzimparmpas, Jessica Hullman, Matt Groh

Acknowledgments

We would like to acknowledge the following people who have contributed to this website and research: Ziv Epstein, Chaz Firestone, Rosalind Picard, Aruna Sankaranarayanan, Nikhil Singh, Dong Young Kim, Andy Lippman for research collaborations; Alicia Guo for excellent design work; Colin Cassidy, Austin Nasso, J L Cauvin, Josh Rosen, and Devin Matthews for their voice acting.