"Fooling Deepfake Detectors"

Deepfakes refer to synthetic media, including images and videos, created using Artificial Intelligence (AI). Deepfakes remain a significant issue despite recent developments in security measures that can detect many of them. Bart Kosko, a professor in the Ming Hsieh Department of Electrical and Computer Engineering, underlined the expected advancement in the generation of deepfakes and the increased difficulty in detecting them as computers and learning algorithms grow more sophisticated. In a new paper from Professor Kosko's neural learning and computational intelligence course, Electrical and Computer Engineering masters students Apurva Gandhi and Shomik Jain showed the vulnerability of deepfake detectors to adversarial perturbations. This article continues to discuss the expected advancement of deepfakes, the results of the study, and the researchers' strategies for improving vulnerable deepfake detectors. 

Homeland Security News Wire reports "Fooling Deepfake Detectors"